This item is only available as the following downloads:
1 EFFICACY AND ECOLOGICAL EFFECTS OF MECHANICAL FUEL TREATMENT S IN PINE FLATWOODS ECOSYSTEMS OF FLORIDA, USA By JESSE K KREYE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLM ENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Jesse K Kreye
3 To our son Raleigh (The "Skootch er ")
4 ACKNOWLEDGMENTS Funding for this project was provided by the Joint Fire Science Program, t he USDA Forest Service, and the American Recovery and Reinvestment Act of 2009. I thank my major advisor, Leda Kobziar, who's continued support and motivation helped me to strive for success and who's professional resolve provide s a template for my future endeavors I thank my committee members, Wendell Cropper, Alan Long, Michelle Mack, and Timothy Martin, for all of their advice and comments. I also thank Wayne Zipperer who's advice and support for this project was very helpful I owe much gratitude to Mike Camp, who made this project a success through all of his hard work in some of the harshest conditions I thank the many others that worked on this project including Nick Bowman, Shannon McGee, Eric Carvalho, Liz Ramirez, Dawn McKinstry. I thank David Godwin for his help especially during prescribed burning. I am very thankful for the support of all of the staff of the Osceola National Forest. I especially thank Peter Myers, Shawn Kinghorn, Byron Hart, and the rest of the Osceola fire crew for the many hours of work they provided to implement our research treatments. I thank Morgan Varner who instilled in me an excitement of scientific inquiry and to continually strive to ask questions, follow them to their resolve, and allow them to reveal new ones. I thank my aunt, Melissa Hennessey, for her inspiration as a firefighter with the USDA Forest Service for so many years. I thank my Mom for her much needed help while I worked on this dissertation and for her many years of support and the rest of m y family who have support ed my endeavors. I thank the many friends who have encouraged me to achiev e my g oals. Lastly, and most importantly, I would like to thank my wife Melissa, for 15 years of supporting my dreams, and our newborn Raleigh, who inspi res me to be a better person.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 2 FU ELBED CHARACTERISTICS FOLLOWING MECHANICAL TREATMENTS OF UNDERSTORY FUEL STRATA IN PINE FLATWOODS ECOSYSTEMS OF FLORIDA, USA ................................ ................................ ................................ ....... 20 Background ................................ ................................ ................................ ............. 20 Methods ................................ ................................ ................................ .................. 21 Study Site ................................ ................................ ................................ ......... 21 Areal Treatment ................................ ................................ ................................ 23 Data A nalysis ................................ ................................ ................................ ... 27 Buffer Treatment Zone ................................ ................................ ..................... 27 Data Analysis ................................ ................................ ................................ ... 29 Results ................................ ................................ ................................ .................... 30 Areal Treatment ................................ ................................ ................................ 30 Buffer Treatment ................................ ................................ .............................. 31 Discussion ................................ ................................ ................................ .............. 34 3 EXPERIMENTAL BURNING IN MASTICATED PALMETTO/GALLBERRY: EFFECTS OF FUEL LOADING AND MOISTURE CONTENT ON FIRE BEHAVIOR AND LETHAL HEATING IN COMPACT LITTER DOMINATED FUELBEDS ................................ ................................ ................................ ............. 55 Background ................................ ................................ ................................ ............. 55 Methods ................................ ................................ ................................ .................. 57 Results ................................ ................................ ................................ .................... 62 Discussion ................................ ................................ ................................ .............. 64 Treatment Effects on Fire Behavior and Above and Belowground Temperatures ................................ ................................ ................................ 66 Saw Palmetto/Gallberry and Other Fuelbed Types Compared ......................... 67 Fireline Intensity ................................ ................................ ............................... 68 4 FIRE BEHAVIOR AND EFFECTS IN MASTICATED PINE FLATWOODS ECOSYSTEMS OF FL ORIDA, USA ................................ ................................ ....... 78
6 Background ................................ ................................ ................................ ............. 78 Methods ................................ ................................ ................................ .................. 81 Field Experimental Burns ................................ ................................ ................. 81 Data Analysis ................................ ................................ ................................ ... 86 Modeled Versus Observed Fire Behavior ................................ ......................... 89 Res ults ................................ ................................ ................................ .................... 90 Winter M+B versus Winter B Treatments ................................ ......................... 90 Winter Versus Summer M+B Treatments ................................ ......................... 93 Modeled versus Observed ................................ ................................ ................ 94 Tree Mortality ................................ ................................ ................................ ... 95 Discussion ................................ ................................ ................................ .............. 96 5 EFFECTS OF MECHANICAL FUEL TREATMENTS AND PRESCRIBED BURNING ON VEGETATION, MICROCLIMATE, AND SOILS IN PINE FLATWOODS ECOSYSTEMS OF FLORIDA, USA ................................ .............. 116 Background ................................ ................................ ................................ ........... 116 Methods ................................ ................................ ................................ ................ 120 Vegetation Dynamics ................................ ................................ ..................... 122 Microclimate and Fuel Moisture Dyna mics ................................ ..................... 125 Decomposition ................................ ................................ ................................ 127 Soil Nutrients ................................ ................................ ................................ .. 129 Results ................................ ................................ ................................ .................. 130 Vegetation Dynamics in the Buffer Area ................................ ......................... 130 Vegetation Dynamics in the Experimental Block Area ................................ .... 132 Microclimate and Moisture Dynamics ................................ ............................. 135 Decomposition ................................ ................................ ................................ 137 Soil Nutrients ................................ ................................ ................................ .. 138 Discussion ................................ ................................ ................................ ............ 138 6 CONCLUSIONS ................................ ................................ ................................ ... 175 LIST OF REFERENCES ................................ ................................ ............................. 179 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 185
7 LIST OF TABLES Table page 2 1 Surface fuel characteristics following mowing in palmetto/gall berry pine flatwoods in northern Florida, USA from destructive sampling. .......................... 39 2 2 Overstory, understory, and surface fuel characteristics of a 500 ha mowing treatment in palmetto/gallberry pine flatwoods of northern Florida, USA. Surface fuels sampled non destructively (planer intercept method). .................. 40 2 3 Overstory characteristics following mowing treatments in three stand types of palm etto/gallberry pine flatwoods of northern Florida, USA. ............................... 41 2 4 Biomass of shrubs, surface fuels, and total (shrubs and surface fuels) following mechanical mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. ................................ ................................ .... 42 2 5 Shrub foliage and stem biomass, shrub height, and shrub density following mechanical mowing of understory shrubs and small trees in pi ne flatwoods of northern Florida, USA. ................................ ................................ ........................ 43 2 6 Biomass of litter and fine woody fuels (1h, 10h, 100h) following mechanical mowing of understory shrubs and small trees in pine flatwoods of north ern Florida, USA. ................................ ................................ ................................ ...... 44 2 7 Biomass of 1000h (sound and rotten) woody fuels and duff following mechanical mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. ................................ ................................ ........................ 45 3 1 Fire behavior characteristics from experimental burning of masticated understory vegetation of southeastern pine flatwoods across fuel loading and fuel moisture content treatments. Marg inal and cell means are listed along with p values from GLM ANOVA. ................................ ................................ ....... 71 3 2 A comparison of observations from this study conducted in constructed fuelbeds of masticated palmetto/gallberry of sou theastern USA pine flatwoods and that of two other studies where burning experiments were conducted with constructed fuelbeds from masticated understory shrub vegetation of western USA forests. ................................ ................................ .... 72 4 1 Weather, overstory, and fuel conditions during experimental burning of masticated (mow+burn) and untreated (burn) stands of palmetto/gallberry pine flatwoods in northern Florida, USA. ................................ .......................... 103 4 2 Fire behavior and effects from burning of masticated (mow+burn) and unmasticated (burn only) palmetto/gallberry pine flatwoods. ............................ 104
8 4 3 Comparison of burning conditions (weather, over story, and fuels) between a summer and winter burn in masticated palmetto/gallberry pine flatwoods of northern Florida, USA. ................................ ................................ ...................... 105 4 4 Fire behavior and effects between summer (July) and winter (Feb) burning of masticated palmetto/gallberry pine flatwoods. ................................ .................. 106 4 5 Number of trees dead or alive across three treatments at one year following burning in palmetto/gallberry pine flatwoods. ................................ ................... 107 5 1 Tree density, basal area, and quadratic mean diameter (QMD) following mowing treatments in pine flatwoods of northern Florida, USA. ....................... 151 5 2 Density and species richness of understory shrubs and small trees following mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. ................................ ................................ ................................ .... 152 5 3 Percent groundcover, by vegetation type, and species richness of shrubs (<0.5 m) and tree saplings (<0.5 m) following mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. ............................. 153 5 4 Tree density, basal area, and quadratic mean diameter (QMD) across experimental treatments following mowing and burning in pine flatwoods of northern Florida, USA. ................................ ................................ ...................... 155 5 5 Density and species richness of understory shrubs and small trees, and percent cover of saw palmetto, across experimental mowing and burning treatments in pine flatwoods of northern Florida, USA. ................................ .... 156 5 6 Percent groundcover, by vegetation type, and species richness of shrubs (<0.5 m) and tree saplings (<0.5 m) across experimental mowing and burning treatments in pine flatwoods of northern Florida, USA. ................................ .... 158 5 7 Soil properties and nutrients across experimental mowing and burning treatments in pine flatwoods of northern Florida, USA. ................................ .... 159
9 LIST OF FIGURES Figure page 2 1 Areal (500 ha) and buffer (60 ha) treatments masticated in palmetto/gallberry pine flatwoods in northern Florida, USA. ................................ ............................ 46 2 2 Fuels and vegetatio n sampling in the areal mowing treatment. .......................... 47 2 3 Fuels and vegetation sampling in the buffer treatment. ................................ ...... 48 2 4 Litter (top) and duff (bottom) mass as a function of depth following mowing treatments in palmetto/gallberry pine flatwoods in northern Florida, USA. Measurement taken just after mowing (left) and one year following mowing (right). ................................ ................................ ................................ ................. 49 2 5 Saw palmetto allometry used for estimation of biomass from non destructive sampling. Frond includes rachis and lamina. ................................ ..................... 50 2 6 Shrubs, surface fuels (litter, 1h, 10h, and 100h fuels), and total fuel (shrub + 1 ) following mowing treatment in 3 stand types (mature, mature/burned (burned 5 yrs prior to mowing ),plantation) of palmetto/gallberry pine flatwoods in northern Florida, USA. (0 time si nce treatment= pre treatment) ................................ ................................ ................... 51 2 7 Shrub foliage and shrub stem biomass, shrub height, and shrub density following mowing treatment in 3 stand types (mature, mature/burned (burned 5 yrs prio r to mowing ),plantation) of palmetto/gallberry pine flatwoods in northern Florida, USA. (0 time since treatment= pre treatment ......................... 52 2 8 Surface fuel components (1h, 10h, 100h, and litter) follo wing mowing treatment in 3 stand types (mature, mature/burned (burned 5 yrs prior to mowing ),plantation) of palmetto/gallberry pine flatwoods in northern Florida, USA. (0 time since treatment= pre treatment) ................................ ................... 53 2 9 Large woody fuels (1000h sound (S) and rotten (R)) and duff biomass following mowing treatment in 3 stand types (mature, mature/burned (burned 5 yrs prior to mowing ),plantation) of palmetto/gallberry pine flatwoods in northern Flori da, USA. (0 time since treatment= pre treatment) ........................ 54 3 1 The relationship 1 between fireline intensity (kJm 1 s 1 ) and flame length (m) during the burning of fuelbeds created from masticated p almetto/gallberry dominated pine flatwoods understory (solid line, R 2 =0.81), compared with Byram's (1959) fireline intensity equation 2 (dotted line). ................................ ..... 73 3 2 The effect of fuel loading on maxim um temperatures reached at the fuelbed surface during the burning of fuelbeds created from masticated palmetto/gallberry dominated pine flatwoods understory. Temperatures
10 differed amongst all three fuel loading treatments using the Tukey Kramer post hoc c omparison of the means. ................................ ................................ .... 74 3 3 The effect of fuel loading on the duration of aboveground surface heating created from masticated palmetto/gallberry dominated pine flatwoods understory. Lethal heating differed amongst the three fuel loading treat ments using the Tukey Kramer post hoc comparison of the means. ............................. 75 3 4 The effect of fuel moisture content (FMC) on soil heating (maximum temperatures) at three soil depths during the burning of fuelbeds created from masticated palmetto/gallberry dominated pine flatwoods understory. ........ 76 3 5 The effect of fuel loading on soil heating at three soil depths during the burning of fuelbed s created from masticated palmetto/gallberry dominated pine flatwoods understory. ................................ ................................ .................. 77 4 1 Experimental mowing and burning treatments in pine flatwoods in northern Florida, USA (Osceola National Forest). Systematic plot locations are indicated. Burn only and mow+burn treatments burned with strip head firing techniques (white arrows indicate fire movement). ................................ .......... 108 4 2 Example of plot locations within the buffer treatments. Sampling within plots were the same for both buffer and experimental block treatments. All trees transects were randomly oriented. ................................ ................................ .... 109 4 3 Fire behavior in experimental mowing and burning treatments in pine flatwoods of northern FL, USA. Burn only treatments were not masticated, mow+burn treatments were masticated 6 months prior to burning. .................. 110 4 4 Fire behavior measurements (rate of spread, above; flame height, below) as a function of shrub cover (left), shrub height (middle), and litter mass (right) during the burn ing of mowed and un mowed experimental treatments in pine flatwoods. ................................ ................................ ................................ ......... 111 4 5 Observed versus predicted fire behavior across burning treatments within mowed (M+B) and un mowed (B) palmetto/gall berry pine flatwoods burned in the winter (Feb) and mowed treatments burned in the summer (M+B summer). Solid line, 1:1 ratio; Dashed line, linear regression. .......................... 112 4 6 Crown scorch (%) versus t ree diameter (DBH) (left) and tree mortality within diameter distributions (right) across burn only (top) and mow+burn (middle) treatments burned in the winter (Feb) and mow+burn treatments burned in the summer (July) (bottom). ................................ ................................ .............. 113 4 7 Tree mortality across individual tree characteristics (height and DBH) and tree damage (crown scorch and bole char height) following burning in
11 masticated and non masticated treatments in palmetto/gallberry pine flatwoods. The height vs DBH graph indicates the only 2 hardwoods in the study (both died) and the only 2 trees that died in the burn only treatment, all other dead trees occurred in the masticated treatment burned in the summer. Trees 1 and 2 are indica ted in both graphs and were both large trees with little crown scorch that died following summer burning following mowing ....... 114 4 8 Distribution of crown scorch (top), bole char height (middle) and percent bole circumference charred at DBH (bottom) across burn only and mow+burn treatments burning in the winter (Feb) and mow+burn treatments burned in the summer (July). ................................ ................................ ............................ 115 5 1 Fuels treatments used for the study of the ecological effects of understory mowing in pine flatwoods of the Osceola National Forest (ONF) in northern Florida, USA. Three treatment areas are shown. 1) a 100 m wide and 6 km (60 ha) buffer masticated ("mowed") in 3 stand types: mature pine (ca. 80 yrs old), mature pine recently burned (5 yrs prior to mowing ), and young pine plantation (28 yrs old); 2) a 500 ha areal treatment (sampling plots exist in mature pine only); and 3) three experimental blocks each with th e following treatments: mow, mow followed by burning, burn only, and control. ................ 162 5 2 Vegetation sampling plots systematically allocated within a fuels treatment buffer on the Osceola National Forest in northern Florida, USA. Plots were located at the center of delineated stand types (mature, mature burned, young plantation). ................................ ................................ ............................. 163 5 3 Tree height and diameter distributions pre and post treatment following mowing in 3 stand types (mature, mature/burned, plantation) in pine flatwoods in northern Florida, USA. ................................ ................................ .. 164 5 4 Density and species richness of understory shrubs and small trees following mechanical mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. ................................ ................................ ...................... 165 5 5 Density by species of understory shrubs (left) and trees <2.5 c m DBH (right). 166 5 6 Groundcover (%), by cover type, and species richness of shrubs (<0.5m in height) and tree seedlings (<0.5 m in height) following mowing in 3 stand types (mature, mature/ burned, plantation) in pine flatwoods in northern Florida, USA. ................................ ................................ ................................ .... 167 5 7 Density and species richness of understory shrubs and small trees across experimental mowing and burning treatments in p ine flatwoods of northern Florida, USA. ................................ ................................ ................................ .... 168 5 8 Percent groundcover, by cover type, and species richness of shrubs (<0.5 m) and tree saplings (<0.5 m) across experimental mowing and burning tr eatments in pine flatwoods in northern Florida, USA. ................................ ..... 170
12 5 9 Average temperature (above) and relative humidity (below) across 3 fuels treatments (burn, mow, mow+burn) and controls up to 17 mon ths following mowing treatments conducted in August 2010. Burning treatments were conducted in Feb 2011, six months following mowing. ................................ ..... 171 5 10 Average soil temperature, at 5 cm depth, across 3 fuels treatments (burn, mow, mow+burn) and controls up to 16 months following mowing treatments conducted in August 2010. Burning treatments were conducted in Feb 2011, six months following mowing. ................................ ................................ ........... 172 5 11 Moisture content (%) of surface litter (left) and live shrub foliage (right) across fuels treatments (mow, mow+burn, burn only), and controls, in mature pine flatwoods of northern Florida, USA. Moisture content sampled every 3 to 4 weeks between June 2011 and March 2012. Inserts indicate moisture content differences by treatment during the driest season. .............................. 173 5 12 Comparison of decomposition of surface litter and surface wood y debris (1h: <0.625 cm; 10h: 0.625 2.54 cm) created from mowing of saw palmetto and gallberry dominated understory of mature pine flatwoods of northern Florida, USA between mowed treatments and un mowed controls. All material collected for decomposition study were derived from understory mowing, however decomposition rates evaluated in un mowed controls was to determine if shrub cover influenced decomposition since shrub recovery following mowing is rapid. No differences in decomposition were detected be tween treatments across any of the fuel types: litter (P=0.249), 1h woody (P=0.386), 10h woody(P=0.438). ................................ ................................ ...... 174
13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Part ial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFICA C Y AND ECOLOGICAL EFFECTS OF MECHANICAL FUEL TREATMENTS IN PINE FLATWOODS ECOSYSTEMS OF FLORIDA, USA By Jesse K Kreye August 2012 Chair: Leda N. Kobziar Major: Forest Reso urces and Conservation Mechanical fuels treatments are being widely used in fire prone ecosystems where fuel loading poses a hazard, yet little research comprehensively examining fuel dynamics, fire behavior, and ecological effects exists, especially in t he southeastern US In order to broaden our understanding of these treatments effects of mechanical mastication ("mowing") were examined in a common pine ecosystem of the southeastern US Coastal P lain, where the post mastication fuel environment is unique among ecosystems where mastication is being employed. Foliar litter dominates surface fuels after understory mastication in palmetto/gallberry pine flatwoods however rapid recovery of shrubs quickly regains control over fire behavior. Treatments were ef fective at reducing flame heights during post treatment burning in these sites, however compact surface fuels were observed to cause long duration heating during laboratory burning. Overstory tree mortality observed following summer burning in these treat ments may have resulted from combustion of the compact surface fuels beneath the shrub layer. Although temperature and humidity at the shrub level were little influenced by treatments, drier surface fuels existed in masticated sites where shrub cover was r educed potentially exacerbating combustibility of th e surface fuel layer.
14 Treatments had little impact on understory vegetation communities or soil nutrients, however reduction in saw palmetto evidenced in this study may alter future groundcover vegetati on as slight increases in grass cover were observed here. The fast recovery of understory vegetation and generally low impact to ecosystem attributes suggest resiliency of these pine flatwoods to mechanical treatment s however their effectiveness at reduci ng fire hazard is likely short lived. Developing tr eatment regimes that utilize prescribed burning to reduce surface fuel loading following mastication will require special attention to treatment timing in order to ensure surface litter consumption, while minimizing potential impacts to the overstory and meeting overall management goals.
15 CHAPTER 1 INTRODUCTION Fire is a dominant ecological process in many ecos ystems worldwide, however maintaining natural fire regimes through active management is often diff icult. Ecosystems vary in frequency, intensity, extent, and predictability of their historical fire regime (Agee 1993). While some ecosystems may go several decades or even centuries without fire, some have developed in the face of frequent fires that burn wit h relatively low intensity Infrequently burned ecosystems will often burn with high intensity fire behavior that results in substantial alterations of ecosystem structure and composition due to years of fuel buildup. Fuel accumulatio n may occur as trees, understory and midstory vegetation and surface debris. When high intensity fire burns in such an ecosystem, it may take decades or centuries to return to pre disturbance structure and composition. In frequently burned ecosystems, h owever, fuel tends to accumulate as understory vegetation (e.g. grasses or shrubs) and surface debris (vegetative detritus) but are burned often enough that large quantities are not accumulated between successive fires. The plants that occur in these eco systems are typically adapted to such a disturbance regime and may even depend on fire for their perpetuation. Therefore, fire adapted species tend to recover quickly following disturbance and thus maintain dominance in these ecosystems. When ecosystems typified by frequent low int ensity fire regimes are subjected to years of fire absence, fire adapted species may be overtaken by fire sensitive species, but also fuel biomass can build to levels where high intensity fire behavior results when fire does occ ur. Prescribed burning is utilized as a mana gement tool to maintain short interval fire frequencies in fire adapted ecosystems and reduce fuel buildup to decrease fire hazard
16 for health and safety of human populations. It is often difficult however, to m aintain frequent enough fire cycles over larg e areas due to logistics and management constraints, especially in the wildland urban interface (WUI) where human population is in close proximi ty to managed ecosystems. I n areas where high fuel buildup has occ urred, it is hazardous to return fire into forest and shrublands where expected fire behavior could pose a risk to the public or cause detrimental damage to the ecosystem. Returning fire to long unburned ecosystems is desirable to mitigate long term fire hazard, but also for ecological restoration purposes. In forest ecosystems where fire frequency has declined through years of fire suppression, and fuel buildup is too hazardous to burn, fuel management techniques are often used to alter fuel structure pri or to reintroduction of fire or as a stand alone treatment option where burning is difficult. In areas where substantial buildup of mid story trees has occurred, treatments are often silvicultural. Thinning may be used to reduce overstory or midstory dens ity and increase average crown base height, reducing the potential for vertical movement of surface fire into forest canopies. Other treatments may target understory shrub fuels by reducing them through mechanical methods which may be used in concert wit h silvicultural treatments. The goals of such treatments include reducing potential fire intensity, lowering the risk of crown or canopy fires, and enhancing ecosystem resistance to future fires (Agee and Skinner 2005) Mastication of understory shrub s an d small tree s is a fuels treatment method that has becom e increasingly us ed across the United States ( US ) (Glitzenstein et al. 2006, Kane et al. 2009, Kobziar et al. 2009, Battaglia et al. 2010, Menges and Gordon 2010) and elsewhere (Molina e t al. 2009, Ca stro et al. 2010) Mastication is a process in
17 which shrubs and small trees are chipped, shredded, or mowed using front end or boom mounted machinery attached to ground based equipment, usually rubber tired or tracked. Mastication machi nery typically con sist of a mastication head with either rotating blades or a rotating cylinder with fixed or flailing cutters. Mastication heads are hydraulically controlled by the operator and thus allow for manipulation of vegetation with little impact to the ground sur face. This is different than methods, such as roller chopping (Watts and Tanner 2006), that u s e weighted drums pulled behind ground equipment to push over and chop understory vegetation, however causing soil damage in the process. Mastication largely impa cts understory vegetation with little impact to ground fuels or overstory trees. Mastication treatments are being us ed in several shrub and forest ecosystems across the US yet much of the research addressing their eco l o gical impact their fuel character istics, or their effectiveness at reducing fire hazard has been conducted in the western US (Busse et al. 2005, Bradley et al. 2006, Hood and Wu 2006, Kane et al. 2009, Kobziar et al. 2009, Vailant et al. 2009, Battaglia et al. 2010, Kreye et al. 2011, Rho ades et al. 2012, Kreye et al. 2012 ). Much of this research has indicated potential consequence s of burning in post treatment surface debris (Busse et al. 2005, Bradley et al. 2006, Knapp et al. 2011, Kreye et al. 2011) as heavy surface fuel loadings resu lt from treatments where fuel loading is not reduced, but only rearranged into compact woody dominated surface fuelbeds (Kane et al. 2009, Kobziar et al. 2009, Battaglia et al. 2010). Reduction in fire behavior from these treatments may come at the cost o f unforeseen ecological impacts.
18 Mastication is being widely employed in the southe aste rn US also and has gained some research attention, however widespread use of these treatments are occurring with little understanding of their effectiveness or impacts. A few studies have begun to compare mastication (mowing) treatments with other fuel treatments such as prescribed burning or roller chopping (Menges and Gordon 2010), however no studies have fully described post treatment fuel characteristics, evalua ted f uel dynamics over time, and determined treatment effectiveness at reducing fire hazard. Pine f latwoods are a common ecosystem in the Coastal P lain of the southeastern US They are typified by an overstory of pines ( Pinus palustris Mill., P. elliottii En gelm., P. taeda L.) with a shrub und erstory. In the lower Coastal P lain flatwoods are dominated by fire resistant P. palustris and P. elliottii in the overstory and by saw palmetto ( Serenoa repens (Bartr.) Small) and gallberry ( Ilex glabra L. (Gray)) shr ubs in the understory. These flatwoods have a frequent fire regime, burning every 3 10 years, with shrubs that recovery quickly following burning being the dominant fu el driving fire behavior Fire management in this ecosystem requires burning at least e very five years, or sooner, to maintain desired fuel characteristics to minimize hazardous fire behavior. Mastication (mowing) treatments are being employed in areas that have gone as little as five years without burning, but are being prioritized in flat woods stands that have gone even longer without fire. While mastication is largely being us ed as a means to alter fuel structure prior to reintroducing fire, their effectiveness at mitigating fire hazard is unknown. And their potential ecological impacts with or without follow up burning, has not been assessed. The uniqueness of this ecosystem regarding its fuel
19 environment (Hough and Albini 1978, McNab et al. 1978) is likely to result in a unique fuel environment when masticated. Mastication has beco me such a widespread fuels treatment method that fully understanding its effectiveness, as well as impacts, across the many ecosystems in which it is being employed is necessary to evaluate its use. Assessing impacts of such treatments on the fuel environ ment, elucidating fire behavior in their resulting fuelbeds, determining their efficacy at fire hazard reduction, and evaluating their ecological impacts will provide a more holistic determination of their effectiveness as a management tool. In order to m ore fully understand mastication as a fuels treatment option in palmetto/gallberry pine flatwoods of the southeastern US the research presented here aimed to evaluate the effects of mastication on the fuel environment, fire behavior, and ecological attrib utes. The objectives of these studies were to 1) describe fuelbed characteristics in masticated stands and e valuate fuel dynamics over time; 2) quantify fuelbed level effects on fire behavior in masticated residues; 3) determine the effect of mastication on fire behavior and effects at the stand scale ; and 4 ) evaluate the effects of mastication and mastication in conjunction with burning on vegetation dynamics, micro climate, fuel moisture regimes, and soil nutrients. Addressing these issues should provid e insight into the effectiveness and impacts of mastication in palmetto/gallberry pine flatwoods and improve our understanding of mastication as a fuels treatment option as a whole.
20 CHAPTER 2 FUELBED CHARACTERIST ICS FOLLOWING MECH ANICAL TREATMENTS OF UNDERSTORY FUEL STRA TA IN PINE FLATWO ODS ECOSYSTEMS OF FLORIDA, USA Background Altering fuel structure in forest and shrub ecosystems has become a common method to mitigate fire hazard in long unburned ecosystems. Mechanical masticat ion (mowing, shredding, chipping, etc.) of understory fuels rearranges shrubs and small trees into compact surface fuels (Hood and Wu 2006, Kane et al. 2009, Kobziar et al. 2009) with the intent to reduce subsequent fire behavior. In order to develop fuel models to aid in the prediction of fire behavior in these treatments, characterizing fuelbeds following mastication across different ecosystems will be important. While recent research has started to describe the post mastication fuel environment, much of this work has been conducted in the western US and has primarily revealed a woody dominated surface fuelbed following treatment (Hood and Wu 2006, Kane et al. 2009, Kobziar et al. 2009, Battaglia et al. 2010). Pine flatwoods of the southeastern US with understories dominated by saw palmetto ( Serenoa repens (Bartr.) Small) and gallberry ( Ilex glabra L. (Gray)) shrubs are unique in regard their fuel c haracteristics (Mcnab et al. 1978 ). Saw p almetto is a shrub palm that grows from horizontal stems and rea ches approximately 2 m i n height Historically, fires were frequent in this ecosystem and understory shrubs typically recover quickly following burning Mastication in this fuel complex will likely result in unique post treatment fuelbeds that may deser ve special attention for fire behavior prediction. Characterizing post mastication fuelbeds in palmetto/gallberry understories will support the creation of
21 fuel models and provide a range of fuelbed characteristics not likely to occur following masticati on in other ecosystems. Mastication or "mowing", of palmetto/gallberry understories in pine flatwoods is being conducted in large scale applications in northern Florida USA to reduce fire hazard during post treatment prescribed burning. Mowing is also being used as a stand alone treatment where burning is difficult in the wildland urban interface, but where altering fuel structure is intended to reduce potential fire behavior during a wildfire. While many shrub species in this ecosystem sprout followi ng aboveground damage and saw palmetto will continue to produce new frond growth following burning, it will be important to understand fuelbed dynamics following treatments to better predict future fire behavior and understand treatment efficacy on mitiga ting fire hazard. The objectives of this study were to 1) characterize surface fuelbeds following the mowing of palmetto/gallberry dominated pine flatwoods and 2) quantify changes in fuels for up to two years following treatment in three stand types: matur e, mature/recently burned, and plantation. Methods Study Site Fuel characteristics were measured in mechanically treated sites on the Osceola National Forest (ONF) in northern peninsular Florida, USA. The ONF encompasses 81,000 ha that occ ur in parts of Columbia, Baker, Bradford, and Hamilton counties. The terrain is generally flat with underlying marine deposited sandy soils. Climate is characterized by hot humid summers with mild winters and most precipitation occurring
22 during summer mo nths from thunderstorms. Dominant vegetation communities on the ONF include mesic and hydric pine flatwoods and cypress hardwood swamps. Mechanical fuels treatments on the ONF were conducted primarily in pine flatwoods communities that have gone unburned f or several years and where fuel accumulations pose a hazard within the wildland urban interface (WUI). Pine flatwoods in this region are dominated by slash pine ( Pinus elliottii var. elliottii (Engelm.)) and/or longleaf pine ( Pinus palustris Mill.) with a n understory comprised primarily of saw palmetto and gallberry Because these systems recover to pre burn fire hazard levels in less than five years (Davis and Cooper 1963), management goals are to burn pinelands on an average three year rotation, althoug h many pine flatwoods areas have not burned in over five years. Challenges to management on the ONF include very large burn units, extensive WUI including major interstate highways, wilderness areas isolated by wetlands, and a history of fire exclusion or excessively long fire return intervals in many locations. Thus, mechanical mowing treatments are being used to create firebreaks, reduce the height of understory fuels for re introduction of prescribed fire, and to reduce fire hazard in areas abutting co mmunities, highways, or large private pine plantations. For this study, fuels were sampled within two mowing treatments in the southwestern portion of the ONF. One, a large contiguous area (50 0 ha) adjacent to Interstate 10 is referred here as the 'area l' treatment, and the other, a 100 m wide, 6 km long buffer treatment (60 ha) is adjacent to private ly owned pine plantations. Each treatment occurred within pine flatwoods ecosystems, however, the areal treatment site was in mature pine (ca. 80 yrs old) flatwoods, while the buffer treatment occurred across three different pine flatwoods stand types: mature (ca. 80 yrs old),
23 mature/burned (ca. 80 yrs old, burned 5 yrs prior to mowing ), and a younger pine plantation (27 yrs old). Areal Treatment T o char acterize fuelbed properties following mowing in pine flatwoods, fuels and vegetation were sampled from 16 plot locations within the 500 ha areal treatment ( Figure 2 1). Plots were allocated using a systematic grid randomly located onto an aerial map of th e treatment zone. A grid format was used such that the distance between all grid line intersections was 400 m. Relative plot locations were systematically located using a grid pattern to better facilitate repeated sampling, however, of all possible grid intersections, 16 were randomly selected as sample locations. In addition, sampling locations were only used that occurred within mature pine stands, i.e. if a randomly selected grid intersection occurred within a wetland, it was not used. Plots were est ablished and vegetation and fuels sampled in January 2010, just prior to mowing to evaluate pre treatment vegetation and fuel loading. Vegetation and fuels were subsequently measured following treatment. At each plot location, all trees were measured within a 201 m 2 (8 m radius) circular plot ( Figure 2 2). Tree diameter at breast height (DBH: measured at 1.37 m above the cm DBH, by species and by tree status (live or dead). sampled within two 4 m 2 rectangular belt transects (18 m) located at 4 m N and S of plot center, respectively, each extending to the 8 m plot radius ( Figure 2.2). Height and basal diameter were measured for al l shrubs. For individual saw palmetto ( Serenoa repens ), fronds were tallied for each individual and an average sized frond was selected for measurement of basal rachis
24 diameter and frond (palm blade and rachis) length. Biomass of shrub woody stems and fo liage were estimated, separately, for the dominant shrub species using published allometric equations (Smith and Brand 1983, Schafer 2010), except for saw palmetto. Saw palmetto biomass was estimated from an allometric equation developed in this study fro m 40 fronds, each collected from 40 different palmetto individuals in an adjacent stand, and regressed against basal rachis diameter and frond length. Gallberry ( Ilex glabra ) and saw palmetto were the most dominant shrub species in this study (Ch 5), howe ver lesser occurrences of Ilex coriacea Vaccinium stamineum V. myrsinites Lyonia lucida L. ferruginea and Myrica cerifera were also present, however species specific allometric equations were not available for all of these species. Allometric equat ions for I. glabra (Smith and Brand 1993) were used for I. glabra and I. coriaceae equations for Vaccinium spp. (Smith and Brand 1993) were used for V. stamineum equations for Myrica pensylvanica (Smith and Brand 1993) were used for M. cerifera and equa tions for Vaccinium scoparium a small statured shrub, were used for V. myrsinites a shrub with similar habit. Because these shrub species were not as abundant in this ecosystem and the respective species used as surrogates were similar in form, biomass estimates across sites are probably reasonable for fuels analysis. Specific allometric equations for Lyonia lucida and L. ferruginea were from Schafer (2010). Herbs, grasses, and vines are a minor component regarding the fuel complex and were not quantif ied for evaluation of fuel dynamics in this study. However, they were assessed in a more complete vegetation analysis in an ecological assessment of treatments in the buffer treatment (Ch 5).
25 Surface fuels were quantified using a non destructiv e planer intercept method (Brown 1974). To estimate coarse (CWD) and fine woody debris (FWD), woody fuels were tallied, by timelag diameter classes, along four 10 m transects extending from 4 m N, S, E, and W, respectively, from plot center, and each orie nted at a random azimuth ( Figure 2.1). FWD include the 1h (<0.635 cm), 10h (0.635 2.54 cm), and 100h (2.54 7.62 cm) timelag fuel classes. 1h and 10h fuels were tallied within the last meter of each transect, away from plot center, and 100h fuels were tallied within the last 2 m. CWD (>7.62 cm) was tallied, and diameter measured, along the entire 10 m transect. CWD was further categorized into two decomposition classes: sound and rotten. 1 ) was estimated from tallies using B rown's (1974) equations and fuel characteristics of palmetto/gallberry pine flatwoods from Hough and Albini (1978). Litter depth and duff depth were measured along each planer intercept transect at the transect origin and at 8 m. Litter mass was then est imated fro m litter depth 3 ) of a 20 yr rough flatwoods site in the longleaf pine (LLP 09) photo series for quantifying natural fuels (Ottmar and Vihnanek 2000). Because duff mass was assumed to not change followin g mowing pre treatment duff mass was estimated from bulk density values measured from destructive sampling following mowing (described below). Following mowing treatment (ca. 2 months), all plots were re sampled using the above methods. To fully descri be post mowing fuelbed characteristics, however, surface fuels (FWD, litter, and duff) were destructively sampled, transported to the laboratory, sorted, oven dried, and weighed. 11 m quadrats were allocated 1m from the end of two randomly selected fuels transects in each plot ( Figure 2 2). All FWD and
26 litter was collected from the entire quadrat and duff collected from a 0.25 x 0.25 m nested quadrat. Woody fuel depth and litter depth were measured at four locations within the quadrat and du ff depths we re measured at four locations within each nested quadrat, prior to the removal of material ( Figure 2 2). Litter and FWD were separated in the laboratory. FWD was subsequently sorted into timelag classes (1,10, and 100h) and further into fractured and non f ractured particles. Fractured particles were those in which a minimum of 50% of the length was physically altered from mowing Litter, samples 'floated' in water for 24 h i ndicated very little mineral soil content (<5% by weight). The transiti on from duff to mineral soil is quite distinct, therefore mineral soil was not removed from duff samples collected from quadrats. At the quadrat level, the relationship between litter mass and average litter depth, as well as the relationship between duff mass and average duff depth were evaluated using linear regression. The resulting linear regression equations were then used to estimating litter and duff mass from depth measureme nts using non destructive planer intercept methods for post masticated sites in the rest of the study Average bulk density was calculated for FWD, litter, and duff. It was assumed that duff biomass was not altered during mowing but that bulk density ma y have increase from machine operations. And since destructive sampling was not conducted prior to treatment, pre treatment bulk density was calculated using average pre treatment duff depth, post treatment duff depth, and post treatment bulk density, ass uming duff mass had not changed. One year following mowing treatment (spring 2011), plots were re sampled using the destructive sampling to determine changes in surface fuel loading and whether litter
27 or duff bulk density changed as surface fuels settl ed over the first year following treatment. One 2525 cm quadrat was randomly located at each plot. Litter and duff depths were measured and debris collected, oven dried, and weighed as was conducted above. Linear regression was also used to determine i f the relationship between litter depth and litter mass, and duff depth and duff mass had changed. Data Analysis Mean, range, and standard deviation were reported for all fuelbed characteristics measured from destructive sampling. Linear regression was used to evaluate the relationships between litter depth and litter mass, as well as duff depths and duff mass, for both post and one year post treatment, from destructive sampling. From non destructive sampling, overstory characteristics (tree density, basal area ( BA ) quadratic mean diameter ( QMD ) tree height, and tree crown base height ( CBH ) ), shrub characteristics (density, height, biomass), biomass of surface fuels (1h, 10h, 100h, 1000h, litter, duff), and fuel depths (FWD, litter, duff) were each compared between pre and post treatment using a repeated measures analysis of variance (ANOVA) with plot as the subject. Tests for differences amongst the means were conducted at the with the Shipiro Wilk and Modified Levene Tests, respectively. As mentioned above, saw palmetto frond biomass was regressed against frond length and frond basal rachis diameter, separately, using linear regression to establish an allometric equation to es timate biomass from non destructive sampling. Buffer Treatment Zone A 100 m wide, 6 km buffer zone was masticated along the southwestern boundary of the ONF adjacent to private pine plantations during the summer of 2009. Shrub
28 vegetation and surface fuels were sampled immediately prior to treatment, and at 2, 8, 16, and 24 months following treatment using the same non destructive sampling methods described above for the areal treatment. Trees were sampled using the same methods as in the areal treatm ent, but were only measured prior to treatment, post treatment, and two years following treatment. The 8 mos sampling period was conducted at the beginning of the growing season (Mar, 2010), 16 mos sampling after the growing season (Oct, 2010), and 24 mos in Aug, 2011. Pre treatment sampling plots were systematically located within the linear buffer and subsequently re sampled following treatment. Allocation of plots within stand types (mature N=12, mature/burned N=9, plantation N=6) were weighted based o n the linear distance of stand types along the buffer. Plots were allocated so that the total number of plots within any one stand type was divisible by three. Plots were spatially arranged in triplets at 15, 45, and 75 m from the buffer edge, but arrang to the edge of the buffer ( Figure 2 3). They were spatially established by locating the center of the stand type unit, t o reduce edge influence from adjacent stand types, and were arranged so that an equal numb er of plots were located on either side of the center of the unit. Shrub biomass was estimate d using the same methods described in the areal treatment. Pre treatment litter and duff mass were estimated from depth measurements using the same procedures as the areal pre treatment estimations. Two and eight month post treatment litter and duff mass were estimated from depth measurements using the regression equations developed from destructive sampling in the areal treatment just after treatment, while l itter and duff mass at 16 and 24 months following
29 treatment were estimated from depth measurements using the regression equations developed from destructive sampling at one year following treatment in the areal site. Data Analysis Overstory tree characte ristics (density, BA, QMD, height, CBH) were compared across stand types (mature, mature burned, plantation) and time since treatment using analysis of variance (ANOVA). Shrub biomass (woody stems and foliage), surface fuel biomass (litter, 1h, 10h, and 1 00h woody), and total fuel biomass (shrub and surface fuel) were compared across stand types and time since treatment using ANOVA. Duff and 1000h fuels were not considered as surface fuel in this particular analysis, but were evaluated separately since th ey contribute to smoldering combustion and not flaming combustion at the fire's front. Shrub characteristics (shrub stem biomass, shrub foliar biomass, shrub height, and shrub density) were each compared across stand types and time since treatment using A NOVA. And biomass of all surface fuels, including duff and 1000h fuels, were each compared across stand type and time since treatment using and the Tukey Kramer post ho c comparison of the means test was used to determine differences amongst groups. Each ANOVA was conducted as a within subjects (repeated measures) analysis with time since treatment as the within subject variable and each plot as the subject. When model assumptions were not met, data were log or square root transformed to meet assumptions.
30 Results Areal Treatment From destructive sampling of surface fuels following treatment, surface fuel 1 with foliar litter accounting for over two thirds of mass, on average (Table 2 1). 1h, 10h, and 100h woody fuels accounted for only 187, 117, and 26 %, respectively. Average litter depth was 5.42.4 cm and litter mass, 1 while average duff depth was 3.62.0 cm and duff mass 41.921.3 1 1 of 1h fue 1 of 10h fuels, were fractured following mowing Only 2 plots had 100h fuels within sampling quadrats, one plot with a fractured particle and one with an unfracture d particle, resulting in a 50% average fracturing of these rare larger fuels. Post mowing litter mass was correlated with litter depth measurements (R 2 =0.93, p<0.001) and a regression equation was developed to estimate mass from depth measurements ( Figure 2 4). One year following treatment, litter ma ss per unit depth was slightly higher than post treatment, however less variation was explained by the regression model (R 2 =0.74, p<0.001; Figure 2 4). Duff mass was well explained by duff depth following treatment (R 2 =0.94, p<0.001) and at one year post treatment (R 2 =0.97, p<0.001), however regression models indicate an almost 20% increase in bulk density one year following treatment ( Figure 2 4). From the 40 saw palmetto fronds collected for allometry, frond mass was best predicted by total frond length (R 2 =0.92, p<0.001) and a regression equation was developed to estimate biomass from non destructive measurements ( Figure 2 5).
31 Pre treatment overstory in the areal treatment consisted of 35839 trees per ha (tph), 18.82.3 m 2 per ha of basal area (BA), and a quadratic mean diameter (QMD) of 25.81.0 cm. Average tree height was 16.70.9 m and crown base height (CBH) was 12.00.8 m (Table 2 2). Following mowing tree density was reduced to 27738 tph (p=0.002), QMD increased to 29.81.2 cm (p=0.002), ave rage tree height increased to 20.70.9 m (p=0.004), and CBH increased to 14.70.7 m (p=0.002). Since only small trees were removed during treatment, BA did not statistically differ following treatment (p=0.577), averaging 18.62.4 m 2 Shrub density (>0.5 m in height) was reduced from 2 2 (p<0.001) following mowing and average height from 1.120.02 to 0.750.14 m (p=0.015). Shrub biomass was 1 prior to mowing 1 afterwa rds (p<0.001). From non destructive sampling (planer intercept method) of surface fuels, biomass of 1 (p=0.022) and 10h fuels 1 (p<0.001) following mowing howe ver 100h and 1000h fuels did not change. Litter depth was reduced from 7.80.8 to 6.00.5 cm 1 (p<0.001). Duff depth was also reduced following treatment (p<0.001), from 5.80.5 to 3.80 .4 cm, but 1 ,however duff mass calculations were developed based on the assumption that it would not be altered by treatment. Average depth of fine woody debris (1h,10h, and 100h) was 7.3 cm and did not change following treatment (p=0.361). Buffer Treatment Mowing in the buffer treatments reduced overstory tree density in all stand types (mature, mature burned, plantation), but only significantly reduced basal area in the
32 mature stands (Table 2 3). While density did not statistically differ between pre and post mowing in the plantation stands, density was lower 2 years following mowing Quadratic mean diameter (QMD) in mature and mature burned stands significantly increased, however QMD was not aff ected by mowing in the plantation. Average tree height increased in both mature stand types following mowing but not in the plantation, however height did statistically increase in plantation stands two years later. In both mature and mature burned stan ds, CBH was increased after treatment, but CBH increased again two years later in the recently burned stands. CBH only differed two years following treatment. Shrub biomass was reduced following treatment in all stand types, but had increased by 16 mont hs ( Figure 2 6, Table 2 4). An i nteraction between time since treatment (TST) and stand type suggested that changes in shrub biomass following treatments differed amongst stand types. Plantations appeared to have less initial shrub biomass than both matu re stands, while mature/burned stands appeared to recover to greater biomass after 16 months than both unburned stand types. Surface fuels increased by about 10 Mg ha 1 in unburned mature stands and plantations, but only increased by 4 Mg ha 1 in the rece ntly burned stands. Total fuel loading (shrubs and surface fuels) did not change in mature/burned stands, however there was evidence of increases in total fuel in the unburned mature stands and especially in plantations. Regarding specific shrub characte ristics, shrub foliage, which should translate into surface litter following mowing was reduced by 2.0, 3.0, and 1.1 Mg ha 1 in mature, mature/burned, and plantation stands, respectively ( Figure 2 7, Table 2 5), while litter increased by 2.2, 2.9, and 5.9 Mg ha 1 ( Figure 2 8, Table 2 6). Shrub foliage
33 increased across all stands by 16 months (Table 2 5). Shrub stems, which should translate into 1 or 10h woody surface fuels (shrub basal diameters were <1.0 cm), were reduced by 3.1, 1.6, and 0.3 Mg ha 1 f ollowing treatment ( Figure 2 7, Table 2 5), while 1h woody fuels increased by 2.8, 1.9, and 2.5 Mg ha 1 and 10h fuels increased by 4.2, 0.4, and 2.9 Mg ha 1 in mature, mature/burned, and plantation stands, respectively ( Figure 2 8, Table 2 6). Therefore, shrub foliage reduction in both mature stand types were similar to litter increases, however more litter was added to surface fuels in plantations than what had occurred in shrub foliage. And while 1h and 10h woody fuel additions to mature/burned stands were close to shrub stem biomass masticated, woody fuel increases in the unburned mature stands and plantation were much higher than accounted for as shrub stem biomass masticated, especially in plantations. Average shrub heights did not differ following treatment across all stands, however shrub density was substantially reduced and recovered to pre treatment densities after 16 months in unburned stands (mature and plantations), but recovery was not as high, by 16 months, in mature/burned stands. As me ntioned above, 1h and 10h surface fuels increased following treatments in all stand types, but 1h fuels were higher than pre treatment values 16 months later ( Figure 2 8, Table 2 6). 10h woody fuels were also higher than pre treatment loading by 16 months but only in unburned mature and plantation stands. 100h surface fuels were not as abundant as 1h and 10h fuels across planer intercepts and did not statistically differ across time since treatment (p=0.500), however these larger fuels were greater in bio mass in the unburned mature and plantation stands compared to the recently burned mature stands. Surface litter increased following treatment, as
34 mentioned above, however mowing in unburned plantation stands resulted in greater litter mass than unburned m ature stands (p=0.006). 1000 h surface fuels, that contribute to smoldering combustion, were rare prior to treatment in all stand types. But, while 1000h sound fuels increased following mowing in all stand types, 1000h rotten fuels did not ( Figure 2 9, T able 2 7). 1000h sound fuels increased more in unburned mature stands than in mature burned stands, and increased even more in plantation stands. While 1000h sound fuels were observed in mature/burned stands after treatment, but not before, they were ver y rare. Duff, which also contributes to smoldering combustion, was not changed just after treatment in both unburned and recently burned mature stands, however duff mass increased following treatment in plantation stands. Duff mass was reduced after 8 mo nths in plantations, while it increased in mature burned stands. Discussion Surface fuelbeds following mowing in these palmetto/gallberry pine flatwoods were dominated by foliar litter, with less proportions of fine woody fuels. This is in contrast to many other post masticated sites that have been studied, where fine woody fuels dominate (Glitzenstein et al. 2006, Kane et al. 2009, Kobziar et al. 2009, Battaglia et al. 2010). F ew studies have addressed mastication in shrub or forest eco systems of the southeastern US especially in pine flatwoods (Menges and Gordon 2010). Of those studies, none fully describe fuelbed characteristics following treatment, but typically address a treatment effect on other attributes. Since pine flatwoods ar e typically burned on a frequent interval, stands that are in need of mechanical treatment from lack of fire may have not burned in as little as five years. Small trees are not abundant, shrubs are
35 not very old, and saw palmetto, a dominant shrub, is prim arily foliar. Therefore, litter dominated surface fuels following mastication is much different than in other ecosystems where treatments occur in older shrublands and forests with substantial under and mid story tree density. Evidence of increased bul k density of litter and duff one year following treatment may be critical to post treatment burning objectives where surface fuel accumulation is desired. Compaction may result in increased moisture retention ( Kreye et al. 2012 ), but also long duration he ating when burned (Busse et al. 2005, Kreye et al. 2011). Meeting management goals when burning in these fuelbeds may require special attention to moisture dynamics in these fuels to ensure desired fuel consumption while minimizing potential effects. Lon g duration heating in compact surface fuels (Kreye et al. 2011) may result in ignition of duff and potential overstory mortality if conditions are dry (Varner et al. 2007). If surface fuels are slow to lose moisture ( Kreye et al. 2012 ), however, desired f uel consumption may not occur even if flammability of shrubs is high enough to carry fire (Gagnon et al. 2010). Effective burning regimes in these novel fuelbeds may require additional knowledge to ensure that management objectives are likely to be met. While shrubs were reduced following mowing in the three stand types studied in the buffer treatment they were recovering quickly as little as 16 months later. Treatment effectiveness in this system may be short lived due to fast recovery of shrub bioma ss on top of the accumulation of surface fuels as a result of treatment. Even shortly after treatments occurred, total fuel that would contribute to flaming combustion (shrubs, litter, and fine woody fuels) was greater in the unburned mature and plantatio n stands in
36 this study. While shrubs ma sticated during treatment translate to surface fuels, even higher total fuel loads in the unburned stands likely result from the smaller trees that were masticated during treatment, but not accounted for as pre treat ment fuels. Understory trees in this study were not considered combustible fuel since they are not primary drivers of fire behavior in this shrub dominated ecosystem (Hough and Albini 1978). Although, when masticated they will likely contribute to surfac e fire behavior as dead woody fuels and leaves are incorporated onto the forest floor. There were less understory trees in the mature stands that had been recently burned and total fuel loading was not increased by mowing This is likely why pre treatmen t shrub stem biomass in the burned stands translated to increases in 1h and 10h woody surface fuels, but more fine woody surface fuels were added to both unburned stand types than what was accounted for in pre treatment shrub stems. Although a window of o pportunity likely exists to conduct post treatment burning prior to shrub recovery, the addition of surface fuels may be an important consideration in evaluating potential ecological consequences when these dense surface fuels burn. Surface litter incr eases following treatment were much larger in plantation stands compared to pre treatment shrub foliage, while they matched well with pre treatment shrub foliage in both the burned and unburned mature stands. While understory trees masticated in both unbu rned stand types may have added to fine woody debris, they may not have contributed as much to litter compared to the shrubs that were masticated. Shrub density and biomass was higher in the unburned mature stands compared to plantations, and recently bur ned stands had even more shrubs than both. Saw palmetto is a dominant shrub in this ecosystem and should contribute heavily to
37 surface litter when masticated since they are primarily foliar. Another potential reason for differences in post treatment litt er accumulation is that litter biomass estimates from depth measurements were calculated using the regression equation developed above. Although litter mass was predicted quite well from post treatment depth measurements, destructive sampling occurred in mature stands, not plantations. If bulk density was lower in plantations, this may account for errors in mass estimation. Mowing equipment was constrained to move linearly in "alleys" between rows of planted pines. Compaction of surface material may hav e been more spatially restricted than in mature stands with less overstory density. Large woody fuels (100h) don't contribute to the flaming front, but may result in undesired fire effects from long duration smoldering. Although rare across these stand t ypes, there were some increases in 1000h sound fuels in this study. Most increases in these larger fuels were in unburned mature stands, and especially in the younger pine plantations, where larger understory trees were masticated. Treatments were such t hat small trees (<20cm DBH) were to be knocked over, but not further masticated after being on the forest floor. Upper portions of downed trees, however, were observed to have been masticated as equipment moved over the surface. This likely attributed to increases in 1h and 10h fuels, while adding to 1000h fuel loading from what remained. Duff is another portion of surface fuels that do es n't contribute to the flaming front, but is an important contributor to smoldering combustion and thus potential fire effects (Varner et al. 2007) and smoke production. It is unlikely that duff mass is affected from mowing even if it is compacted or rearranged. The increase in duff mass just after mowing in the plantation stands may have resulted from error associated with using the duff mass
38 equation developed from the mature stands in the areal treatment above, however duff mass was then reduced at 8 months following treatment. The additional compaction of duff observed at one year following treatment in the areal tr eatment may have also occurred in the plantation stand, as evidenced by a decrease in mass at 8 months. Nonetheless it is unclear why such differences between stand types occurred and the use of the duff estimation equation in the pine plantation may not be appropriate. This study revealed that post mastication surface fuels in pine flatwoods are unique in their high proportion of litter, something not observed with mastication treatments in other ecosystems and their fast recovery of shrub fuels. Whil e shrubs are reduced following mowing the effectiveness of treatments at altering fire behavior may be short lived and follow up prescribed burning to reduce fuel loads or reintroduce fire to long unburned stands will likely need to occur soon following m owing The addition of surface fuels, however, especially in unburned pine flatwoods, may present fire managers with potential problems if burning in these compact su rface fuels results in damage to fine roots or basal cambial tissue of trees (Varner et al. 2007, O'Brien et al. 2010 a ). Considerations regarding surface, duff, and soil moisture will need to be taken into account if prescribed burning is utilized as a follow up treatment with the goals of consuming surface fuels created from mowing While this study provides insight into the dynamics of fuel characteristics following mowing in palmetto/gallberry pine flatwoods of the southeastern US further research will be needed to elucidate how these fuel treatments burn and what potential ecological co nsequences may ensue from their use.
39 Table 2 1 Surface fuel characteristics following mowing in palmetto/gallberry pine flatwoods in northern Florida, USA from destructive sampling. Fuel Load Fuel Depth Fractured a Fuelbed Proportion b ( Mg ha 1 ) ( cm ) ( % ) ( % ) range mean ( sd) range mean ( sd) range mean ( sd) range mean ( sd) Litter 5.6 24.4 12.6 (5.5) 2.4 10.9 5.4 (2.4) na 40 88 69 (13) 1h 1.4 6.0 3.1 (1.2) 6 33 20 (8) 7 29 18 (7) 10h 0.6 6.3 2.1 (1.5) 3.0 12.8 c 7.4 (3.0) c 0 65 25 (20) 4 32 11 (7) 100h 0.0 5.9 0.4 (1.5) 0 100 50 (71) 0 24 2 (6) Total 9.6 35.6 18.2 (6.6) 3.9 13.2 8.1 (2.8) na 100 100 Duff 15.0 98.2 41.9 (21.3) 1.0 8.6 3.6 (2.0) na na a Percent of woody fuels (1,10, and 100h), by weight, that has been fractured at l east 50% of its particle length, b Proportion, by mass, of the total fuelbed associated with flaming co mbustion(does not include duff), c Depth of all fine woody debris (1h, 10h, and 100h)
40 Table 2 2 Overstory, understory, and surface fuel characteristics o f a 500 ha mowing treatment in palmetto/gallberry pine flatwoods of northern Florida, USA. Surface fuels sampled non destructively (planer intercept method). Trees Shrubs a Density BA QMD Height CBH Density Heigh t Biomass trees h a 1 m 2 ha 1 cm m m 2 m 1 Pre Treatment 358 (39) A 18.8 (2.3) A 25.8 (1.0) A 16.7 (0.9) A 12.0 (0.8) A 4.2 (0.5) A 1.12 (0.02) A 3.68 (0.49) A Post Treatment 277 (38) A 18.6 (2.4) A 29.8 (1.2) B 20.7 (0.9) B 14.7 (0.7) B 0.6 (0.2) B 0.75 (0.14) B 0.24 (0.08) B Surface Fuel Loading 1h 10h 100h 1000h S 1000h R Litter Duff -------------------------------------1 ----------------------------------Pre Treatment 1.7 (0.3) A 1.4 (0.1) A 0.3 (0.1) A 0.3 (0.3) A 0.2 (0.2) A 9.0 (0.9) A 42. 0 (3.6) A Post Treatment 2.7 (0.5) B 3.1 (0.5) B 0.6 (0.3) A 0.4 (0.2) A 0.3 (0.2) A 13.4 (1.2) B 42.0 (4.3) A Fuel Depth FWD b Litter Duff --------------cm --------------Pre Treatment 7.2 (1.7) A 7.8 (0.8) A 5.8 (0.5) A Post Treatment 7.3 (0.9) A 6.0 (0.5) B 3.8 (0.4) B a Shrubs >0.5 m in height b Fine woody debris (1h, 10h, and 100h fuels) Note: Va lues sharing letters within columns are not statistically different (
41 Table 2 3 Overstory characteristics following mowing treatments in three stand types of palmetto/gallberry pine flatwoods of northern Florida, USA. Stand Type mature mature burned plantation Stand Type TST a Stand Type TST p valu e Tree Density -------------t rees h a 1 ------------<0.001 <0.001 <0.001 Pre Treatment 941 (179) A 365 (36) A 1120 (185) A Post Treatment 327 (58) B 216 (30) B 804 (82) AB 2yrs Post Treatment 290 (46) B 216 (30) B 713 (71) B Basal Area -----------m 2 ha 1 -----------0.029 <0.001 0.043 Pre Treatment 28.3 (3.3) A 17.9 (2.2) A 34.0 (5.9) A Post Treatment 23.2 (2.9) B 17.3 (2.4) A 27.5 (2.2) A 2yrs Post Treatment 23.4 (2.8) B 18.2 (2.3) A 26.3 (2.5) A QMD -------------cm -------------0.004 <0.001 < 0.001 Pre Treatment 21.8 (1.7) A 25.6 (2.1) A 20.7 (0.4) A Post Treatment 32.2 (2.0) B 32.8 (2.2) B 21.0 (0.5) A 2yrs Post Treatment 33.6 (1.7) B 33.9 (2.2) B 21.8 (0.6) A Height ---------------m --------------0.211 <0.001 <0.001 Pre Treatment 12.9 (0.8) A 16.2 (1.4) A 18.9 (0.4) A Post Treatment 20.3 (1.4) B 21.7 (0.4) B 19.0 (0.2) A 2yrs Post Treatment 22.0 (1.1) B 22.8 (0.5) B 21.9 (0.5) B CBH ---------------m --------------0.158 <0.001 <0.001 Pre Treatment 8.3 (0.7) A 10.5 (0.8) A 13.6 (0.5) A Post Treatment 13.3 (1.1) B 13.2 (0.4) B 13.6 (0.3) A 2yrs Post Treatment 14.8 (1.1) B 15.7 (0.4) C 15.7 (0.4) B a Time Since Treatment Note: Va lues sharing letters within columns are not statistically different (Tukey
42 Table 2 4 Biomass of shrubs, surface fuels, and total (shrubs and surface fuels) following mechanical mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. Stand Type Mature Mature/Burned Plantation Stand Type TST a Stand Typ e TST --------------------Mg ha 1 ----------------------------------------------p value -----------------------Shrubs 0.098 <0.001 0.0 0 8 Pre Treatment 5.5 (1.4) A 5.0 (0.9) A 1.6 (0.5) A 2 months 0.6 (0.3) B 0.5 (0.2) B 0.2 (0.1) B 8 months 0.4 (0.2) B 0.7 (0.2) B 0.5 (0.3) B 16 months 0.8 (0.2) C 2.3 (0.7) C 1.1 (0.5) C 24 months 1.3 (0.3) C 2.1 (0.5) C 0.9 (0.3) C Surface Fuels b <0.001 <0.001 0. 1 0 3 Pre Treatment 11.1 (1.4) A 13.2 (1.4) A 13.9 (1.1) A 2 months 20.7 (1.7) B 17.1 (1.7) B 23.1 (2.7) B 8 months 15.9 (0.9 ) B 14.0 (0.7) B 22.1 (2.3) B 16 months 15.8 (1.3) B 17.0 (1.8) B 24.9 (2.8) B 24 months 16.2 (0.8) B 15.7 (1.6) B 24.1 (1.6) B Total Fuel c 0.007 0.004 0.009 Pre Treatment 16.6 (1.6) A 18.2 (1.5) A 15.5 (1.3) A 2 months 21.1 (1.9) B 17.6 (1.8) A 23.3 (2.8) AB 8 months 16.1 (0.9) A 14.7 (0.9) A 22.5 (2.3) AB 16 months 16.7 (1.3) AB 19.3 (2.2) A 26.0 (2.9) B 24 months 17.4 (0.8) AB 17.8 (1.8) A 25.0 (1.6) B a Time Since Treatment, b includes litter, 1h, 10h, and 100h fuels; c shrubs and surface fu els Note: Va lues sharing letters within columns are not statistically different (Tukey
43 Table 2 5 Shrub foliage and stem biomass, shrub height, and shrub density following mechanical mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. Stand Type Mature Mature/Burned Plantation St and Type TST a Stand Type TST --------------------Mg ha 1 ----------------------------------------p value ---------------------Shrub Foliage 0.049 <0.001 0.306 Pre Treatment 2.4 (0.5) A 3.2 (0.6) A 1.2 (0.4) A 2 months 0.4 (0.2) B 0.2 (0.1) B 0.1 ( 0.1) B 8 months 0.3 (0.2) B 0.4 (0.1) B 0.2 (0.1) B 16 months 0.5 (0.2) C 1.5 (0.4) C 0.6 (0.2) C 24 months 0.9 (0.2) C 1.4 (0.4) C 0.5 (0.1) C Shrub Stems 0.308 <0.001 0.002 Pre Treatment 3.1 (1.1) A 1.8 (0.6) A 0.4 (0.2) A 2 months 0.2 (0.1) B 0.2 (0.2) B 0.1 (0.1) B 8 months 0.1 (0.0) B 0.3 (0.1) BC 0.3 (0.2) BC 16 months 0.3 (0.1) B 0.8 (0.3) C 0.5 (0.3) AC 24 months 0.4 (0.1) B 0.7 (0.2) C 0.4 (0.2) AC Shrub Height ------------------------m -------------------------0.347 0.078 0.788 P re Treatment 0.86 (0.10) A 0.79 (0.08) A 1.00 (0.08) A 2 months 0.72 (0.08) A 0.67 (0.03) A 0.81 (0.12) A 8 months 0.62 (0.14) A 0.73 (0.04) A 0.79 (0.17) A 16 months 0.68 (0.08) A 0.78 (0.03) A 0.80 (0.02) A 24 months 0.89 (0.08) A 0.76 (0.04) A 0.86 (0 .03) A Shrub Density ---------------individuals m 2 ----------------0.018 <0.001 0.018 Pre Treatment 4.9 (0.8) A 13.3 (3.4) AC 2.3 (0.6) A 2 months 0.5 (0.1) B 1.1 (0.3) B 0.8 (0.2) B 8 months 0.6 (0.2) B 1.5 (0.3) BC 0.6 (0.3) B 16 months 3.7 (0.9) A 5.2 (0.7 BC 2.3 (0.5) A 24 months 4.6 ( 1.1) A 7.6 (2.1) C 2.7 (0.4) A a Time Since Treatment Note: Va lues sharing letters within columns are not statistically different (Tukey
44 Table 2 6 Biomass of litter and fine woody fuels (1h, 10h, 100h) following mechanical mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. Stand Type Mature Mature/Burned Plantation Stand Type TST a Stand Type TST --------------------Mg ha 1 ---------------------------------------------p value -------------------------1h woody 0.546 <0.001 0.090 Pre Treatment 0.4 (0.1) A 0.4 (0.1) A 0.2 (0.0) A 2 months 3.2 (0.6) B 2.3 (0.3) B 2.7 (0.4) B 8 months 1.6 (0.4) C 1.5 (0.2) C 1.9 (0.4) C 16 months 0.8 (0.1) D 0.8 (0.1) D 1.5 (0.2) D 24 months 0.6 (0.1 ) D 0.9 (0.1) D 0.8 (0.0) D 10h woody <0.001 <0.001 <0.001 Pre Treatment 1.1 (0.3) A 3.0 (0.7) A 2.5 (0.7) A 2 months 5.3 (0.9) B 2.6 (0.4) A 5.4 (1.0) B 8 months 3.0 (0.5) A 1.9 (0.3) A 6.3 (1.4) B 16 months 2.5 (0.3) A 1.9 (0.3) A 5.6 (0.9) B 24 months 2.7 (0.4) A 1.9 (0.3) A 5.3 (0.8) B 100h woody <0.001 0.500 0.060 Pre Treatment 1.9 (1.2) 1.0 (0.5) 3.5 (0.9) 2 months 2.8 (0.8) 0.4 (0.3) 1.3 (0.6) 8 months 2.0 (0.5) 0 .0 (0 .0 ) 2.4 (1.5) 16 months 2.5 (0.6) 0 .0 (0 .0 ) 3.1 (1.5) 24 months 1.6 (0.5) 1.8 (1.5) 2.7 (0.8) Litter 0.006 <0.001 0.201 Pre Treatment 7.7 (0.5) A 8.9 (0.9) A 7.7 (0.9) A 2 months 9.5 (0.9) BC 11.8 (1.6) BC 13.6 (2.7) BC 8 months 8.5 (0.9) B 10.6 (0.7) B 11.4 (0.9) B 16 months 9.8 (0.6) C 14.3 (1. 8) C 14.7 (1.4) C 24 months 10.9 (0.6) C 11.1 (0.8) C 15.3 (1.0) C a Time Since Treatment Note: Va lues sharing letters within columns are not statistically different (Tukey
45 Table 2 7. Biomass of 1000h (sound and rotten) woody fuels and duff following mechanical mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. Stand Type Mature Mature/Burned Plantation Stand Type TST a Stand Type TST --------------------Mg ha 1 ---------------------------------------p value ---------------------1000h Sound 0.002 <0.001 0.138 Pre Treatment 0.2 (0.2) A 0 .0 (0 .0 ) A 0 .0 (0 .0 ) A 2 months 1.9 (0.7) B 0.4 (0.3) B 4.9 (2.1) B 8 months 2.6 (1.0) B 0.4 (0.2) B 3.1 (0.8) B 16 months 2.5 (1.2) B 0.2 (0.2) B 3 .4 (1.3) B 24 months 2.7 (1.3) B 0.7 (0.6) B 6.3 (2.1) B 1000h Rotten 0.874 0.269 0.755 Pre Treatment 1.0 (1.0) A 0 .0 (0 .0 ) A 0.7 (0.5) A 2 months 0 .0 (0 .0 ) A 0.2 (0.2) A 0.2 (0.2) A 8 months 0 .0 (0 .0 ) A 0 .0 (0 .0 ) A 0 .0 (0 .0 ) A 16 months 0 .0 (0 0 ) A 0 .0 (0 .0 ) A 0.1 (0.1) A 24 months 0 .0 (0 .0 ) A 0.2 (0.2) A 0 .0 (0 .0 ) A Duff 0.179 0.048 0.004 Pre Treatment 62.1 (10.1) A 34.3 (5.5) A 36.3 (7.2) A 2 months 56.5 (5.8) A 37.9 (2.9) A 67.3 (7.0) B 8 months 64.3 (9.2) A 53.1 (6.6) B 50.9 (2.9) AB 16 months 57.6 (8.2) A 51.1 (5.1) B 57.8 (9.9) AB 24 months 48.5 (6.9) A 46.6 (4.4) AB 61.2 (3.1) B a Time Since Treatment Note: Value s sharing letters within columns are not statistically different (Tukey
46 Figure 2 1 Areal (500 ha) and buffer (60 ha) treatments masticated in palmetto/gallberry pine flatwoods in northern Florida, USA.
47 Figure 2 2 Fuels and veg etation sampling in the areal mowing treatment.
48 Figure 2 3 Fuels and vegetation sampling in the buffer treatment.
49 Figure 2 4 Litter (top) and duff (bottom) mass as a function of depth following mowing treatments in palmetto/gallberry pine flatwoods in northern Florida, USA. Measurement taken just after mowing (left) and one year following mowing (right).
50 Figure 2 5 Saw palmetto allometry used for estimation of biomass from non destructive sampling. Frond includes rachis and lamina.
51 Figure 2 6 1 ) following mowing treatment in 3 stand types (mature, mature/burned (burned 5 yrs prior to mowing ),plantation) of palmetto/gallbe rry pine flatwoods in northern Florida, USA. (0 time since treatment= pre treatment)
52 Figure 2 7 Shrub foliage and shrub stem biomass, shrub height, and shrub density following mowing treatment in 3 stand types (mature, mature/burned (burned 5 yrs prior to mowing ),plantation) of palmetto/gallberry pine flatwoods in northern Florida, USA. (0 time since treatment= pre treatment
53 Figure 2 8 Surface fuel components (1h, 10h, 100h, and litter) following mowing treatment in 3 stand types (mature, mature/burned (burned 5 yrs prior to mowing ),plantation) of palmetto/gallberry pine flatwoods in northern Florida, USA. (0 time since treatment= pre treatment)
54 Figure 2 9 Large woody fuels (1000h sound (S) and rotten (R)) and duff bi omass following mowing treatment in 3 stand types (mature, mature/burned (burned 5 yrs prior to mowing ),plantation) of palmetto/gallberry pine flatwoods in northern Florida, USA. (0 time since treatment= pre treatment)
55 CHAPTER 3 EXPERIMENTAL BURNING IN MASTICATED PALMETTO/ GALLBERRY: EFFECTS OF FUEL LOADING AND MOISTURE CONTENT ON FIRE BEHAVIOR AND LE THAL HEATING IN COMPACT L ITTER DOMINATED FUELBEDS Background Mechanical manipulation of forest and shrubland fuels has become an increasingly common approac h to mitigate potential hazards associated with wildfire. Mechanical treatments are frequently utilized within the wildland urban interface (WUI) where risk to life and property are greatest, but are also employed as a restoration tool in fire dependant e cosystems where historical fire regimes have been altered. Such treatments play the role of a fire surrogate in areas where prescribed burn implementation is difficult. Mastication differs from other fuels reduction methods, such as roller chopping, becau se ground fuels and soils are not impacted (Glitzenstein et al. 2006). As such treatments are increasingly being implemented, it is important to fully understand their impacts on potential fire behavior and fire effects. Fuels treatments may be used in concert with prescribed burning or as a stand alone management option. In conjunction with prescribed burning, mastication is used to alter fuel structure prior to implementing fire. The mastication of shrub and small tree understories is intended to re duce flame lengths, thus reducing potential overstory tree mortality and increasing control during burning operations. The conversion of live shrubs and small trees into dead surface fuels can reduce the vertical continuity of fuel strata and the overall fuelbed depth, but increases fuelbed bulk density. If left on site, fuels are only rearranged, with no immediate reduction in total fuel loading (Kobziar et al. 2009, Vaillant et al. 2009). Surface fuel loading is increased, especially in the small diame ter classes (Kane et al. 2009, Kobziar et al. 2009). Fire behavior in densely
56 compacted fuelbeds following mastication has been shown to result in aboveground (Kreye et al. 2011) and belowground (Busse et al. 2005) heating that may conflict with managemen t objectives and have unf oreseen ecological consequences Studies have begun to describe fuel conditions following mastication and to quantify fire behavior in treated sites (Glitzenstein et al. 2005, Bradley et al. 2006, Knapp et al. 2006, Kane et al. 2009, Kobziar et al. 2009). Negative effects on both tree mortality (Bradley et al. 2006) and crown damage (Knapp et al. 2006) have been documented after burning in masticated sites Laboratory studies have also reported that burning of masticated fuelb eds may result in long duration heating both within the soil (Busse et al. 2005) and above the ground (Kreye et al. 2011). Most of the existing mastication research has been conducted in the western US Mastication ("mowing") treatments are being incre asingly employed in the fla twoods forests of the southern Coastal P lain, but their effects have not been examined. Flatwoods forests are a fire dependant ecosystem typified by a historical high frequency, low intensity fire regime (Abrahamson and Hartnett 1990). The understory component is comprised mostly of gallberry ( Ilex glabra (Bartr.) Small) and saw palmetto ( Serenoa repens (L.) Gray) and when masticated, results in high ace of the forest floor. While previous research has found moderate to high proportions, by weight, of fine woody particles in surface fuels of mastication treatments (89%, Glitzenstein et al. 2006; 87%, Kane et al. 2009; 51%, Kobziar et al. 2009), fuelbe ds resulting from mastication in gallberry/palmetto flatwoods are composed of both foliar litter and wood particles with foliar litter being dominant (66%: Kreye unpublished data).
57 This study had two objectives: 1) to evaluate the effects of fuel loa ding and fuel moisture content (FMC) on fire behavior characteristics from the burning of fuelbeds created from masticated understories in southeastern pine flatwoods and 2) to evaluate the effects of fuel loading and fuel moisture content (FMC) on above and below ground heating during the burning of these fuels. To address our first objective we tested the hypotheses that maximum flame length, forward rate of spread (ROS) of the flaming front, percent fuel consumption, and fireline intensity would differ across three fuel load (10, 20, and 30 Mg/ha) and two fuel moisture content (FMC) treatments (low and moderate). We expected flame length and fireline intensity to increase with higher fuel loads, due to higher potential energy available for combustion. We expected the same results in drier fuelbeds, due to a faster rate of combustion as measured by ROS. We also determined the relationship between fireline intensity and flame length and compared it with Byram's (1959) fireline intensity equation. To add ress above and belowground heating, we tested the hypotheses that maximum temperature and duration of lethal temperatures would differ in relation to fuel load and FMC. We expected maximum temperatures and duration of lethal heating to increase with highe r fuel loading, due to our expected increase in fireline intensity, but that all heating would decrease with soil depth. Methods Masticated fuels were collected from a pine flatwoods site in the Osceola National Forest in north central Florida. The site w as dominated by longleaf pine ( Pinus palustris Mill.) and slash pine ( Pinus ellio t tii Engelm.) in the overstory, and by saw palmetto and gallberry in the understory prior to mowing conducted in April 2010. Understory shrubs and small trees (<20cm) were ma sticated using a front end mounted masticator
58 attached to a Gyrotrack. Surface fuels were collected approximately 2 3 weeks following mowing 10 days. To conduct experimental burning, 18 fuelbeds were created from the collected fuel and subsequently burned in May 2010 at the University of Florida Austin Cary Memorial Forest approximately 16 km northeast of G ainesville, FL, USA. Burns were conducted during the typical wildfire season and under warm (28 (46 63% relative humidity), and light wind (0.3 1.8 ms 1 ) conditions. Fuelbeds were burned under three fuel loading treatments (10, 20, and 30 Mgha 1 ) and two fuel moisture content (FMC) treatments (low and moderate) in a 3x2 factorial experimental design, replicated three times. To create two FMC treatments, half of the fuel remained in the drying oven, while the other half was stored in a greenhouse until burning experiments were conducted. Temperature and humidity were not precisely controlled in the greenhouse, but conditions were cooler and wetter than the oven. Three fuel samples were taken from each fuelbed to estimate FMC prior to ignition using the oven dry method. Fuelbeds were created within 4 m diameter circular rings, constructed of 15 cm aluminum flashing, located in a treeless opening within a pine flatwoods forest similar to methods used by Zipperer et al. (2007) Sur face vegetation (primarily grass) was removed prior to loading. Soils on which fuelbeds were created were somewhat poorly drained Grossarenic Paleudults of marine origin with fine sands in the upper 20 cm. The 4 m diameter rings were loaded with 12.6, 25. 1, or 37.7 kg of masticated fuel to create 10, 20, and 30 Mgha 1 fuel loading treatments, respectively. For low FMC treatments, fuel from the oven was taken to the site and kept in a covered truck bed until loading of
59 each ring plot directly prior to bur ning. To create each fuelbed, fuel was placed within the ring and spread out to reach uniform loading. Fuel was tamped down to mimic compact fuelbeds observed in the field as a result of mowing machinery. Low FMC treatments were burned immediately follo wing loading and sensor setup. Each replicate was loaded and then subsequently burned prior to loading the next replicate burn so that fuelbeds would remain as dry as possible prior to each burn. For moderate FMC treatments, fuel that had been stored in a greenhouse for several days were used to create each of nine fuelbeds across the three fuel loading treatments. Fuelbeds were setup, water was applied with a hose, and subsequently covered with plastic for adsorption of moisture into fuel particles for approximately 18 hrs prior to burning. Individual fuelbeds (burn replicates) remained covered until prepared for burning. Thermocouples were located within ring plots to record temperatures above and below ground during combustion. At the center of each ring plot, three 30 AWG Type K PFA insulated thermocouple wires (Omega Engineering, Stamford, CT, USA) were buried to depths of 2, 5, and 8 cm below the soil surface. Wires were buried horizontally in orientation and exposed junctions were inserted appro ximately 10 cm into an exposed vertical soil profile, approximately 10 cm deep, to reduce soil disturbance at the location of temperature measurement. The cavity created for soil thermocouples was then backfilled. At the fuelbed surface, three high temp erature Type K Thermocouple probes (Omega Eng., Stamford, CT, USA) were placed at 1, 2, and 3 m from the ring plot edge, and perpendicular to the anticipated flame front, to record surface temperatures during burning. All thermocouples were connected to an OMB DAQ 55 datalogger (Omega Engineering, Stamford, CT, USA) and temperatures were
60 recorded every 3 seconds. Six poles, with alternating 20 cm black and white measurement markings, were placed at 1.0, 1.5, 2.0, 2.5, 3.0, and 3.5 m from the ignition edge o f the ring plots and perpendicular to the anticipated flame front to estimate flame heights and to estimate the fire's rate of spread. Litter pins (4 ea) were placed at the four cardinal directions, and 1.0 m from the ring's edge, with the top of the litte r pin placed at the fuelbed surface to measure pre and post burn fuelbed depth. Wind speed, air temperature, and relative humidity were measured prior to ignition for each burn. To ignite each fuelbed, a line of fire was initiated perpendicular to th e anticipated spread of the fire at 0.5 m from the edge of the ring using a drip torch. All burns were video recorded from a horizontal position 4.0 m from each ring plot and at 1.5 m above the ground. Maximum flame height was visually estimated at each height pole as the flaming front passed and the time of arrival of the flaming front from ignition was recorded. Maximum flame height was measured as the height of flame that was continuous from the fuel surface, i.e. not including flickering flames detac hed from the main flaming front. Flame length was determined by dividing observed flame heights by the sine of the average flame angle (Rothermel and Deeming 1980). Following combustion, depth of fuel consumed was measured at the four litter pins. Pr oportion consumed was calculated as the depth of fuel consumed divided by pre burn fuelbed depth. Rate of spread (ROS) was calculated as the average ROS between each height pole. Fireline intensity was also determined by multiplying the forward rate of s pread (ms 1 ) of each burn by the proportion of fuel consumed, the fuel load (kgm 2 ), and fuel heat content (kJkg 1 ) (Van Wagner 1973). A heat content value
61 of 19,678 kJ/kg was used from of composite of low heat content values, assuming latent heat of vaporization as a loss, of saw palmetto, gallberry, and a mixture of other pine flatwoods shrubs (Hough and Albini 1978) and adjusted for a 20% nominal energy loss due to radiation (Nelson and Adkins 1986). To evaluate the effects of fuel load and FMC o n the burning of masticated fuelbeds, maximum flame height, ROS, consumption, and fireline intensity were compared across both fuel loading and FMC treatments using a GLM analysis of variance. Both main effects and their interaction were tested at the 0.0 5 alpha level. Model assumptions of normality and equal variance were validated using the Shipiro Wilk Test and the Modified Levene Equal Variance Test, respectively. Where a significant effect of fuel load was detected the Tukey Kramer Test was used to determine differences amongst treatment mean s. The relationship between flame length and fireline intensity was modeled using non linear regression assuming an exponential increase in fireline intensity with flame length (Byram 1959). To evaluate abo ve ground heating at the fuelbed surface, where the potential for basal damage to trees is most likely in these compact fuelbeds, we tested the effects of fuel loading and FMC treatments on both maximum surface temperatures and the duration of lethal heati ng using general linear model procedures (SAS version 9.2, SAS Institute Inc., Cary, NC, USA). Maximum temperatures were compared across low and moderate FMC and the three fuel loadings (10, 20, and 30 Mgha 1 ), and their interaction, to determine how FMC and fuel load influence heating near the fuelbed fuel loading and their interactions. Thermocouple location s within burns were treated as
62 subsamples and were nested within t reatments when testing for main effects. Effects were tested at the 0.05 alpha level and GLM model assumptions were validated as described above. To evaluate soil heating we tested the effects of fuel loading and FMC treatments on soil temperatures acro ss the three soil depths (2, 5, and 8cm) using a GLM analysis of variance. Temperatures were compared across soil depth, FMC, and fuel loading as well as all interactions to determine how FMC and fuel load influence heating at shallow soil depths. Pre bu rn soil temperatures were used as covariates in analysis. Effects were tested at the 0.05 alpha level and GLM model assumptions were validated as above. Results The manipulation of fuelbeds resulted in a low (8.90.6%) and a moderate (12.92.0%) fuel moi sture content (FMC) treatment. One fuelbed was burned with a FMC of 35.6% and was therefore excluded from analysis. Air temperature ranged from 27 .8 to did not differ across FMC or fuel load treatments and were not significant covariates in any analysis. Wind speed during burning was light (0.3 to 1.8 ms 1 ) and did not differ across treatments n or was it a significant covariate in any analysis. Both fuel moisture content (FMC) and fuel loading were significant factors affecting flame lengths and fireline intensity during the burning of fuelbeds created from masticated pine fl atwoods understory (Table 3 1). Flame lengths increased approx. two thirds under the drier versus the wetter (111 cm and 67 cm, respectively) FMC treatment (P=0.001). Flame lengths also increased directly with fuel loading (P<0.001) by approximately 0.5 m per 10 Mgha 1 increase in load, where the 10, 20, and 30
63 Mgha 1 treatments burned with 49, 91, and 140 cm flame lengths respectively. There was no interaction effect (P=0.808) between FMC and fuel loading on flame lengths. Fireline intensity was gre ater in the drier (593 kJm 1 s 1 ) versus the wetter (317 kJm 1 s 1 ) FMC treatments (P=0.029) and also differed across fuel loading (P=0.003), but only between the lowest (10 Mg/ha: 183 kJm 1 s 1 ) and the highest (30 Mg/ha: 773 kJm 1 s 1 ) fuel loading (Table 3 1) There was no interaction effect (P=0.758) between FMC and fuel load on fireline intensity. The relationship between fireline intensity (kJm 1 s 1 ) and flame length (m) determined from non linear regression was where I is fireline intensity (kJm 1 s 1 ) and FL is flame length in meters ( Figure 3 1, R 2 =0.81). Fireline intensity was higher, across our measured flame lengths, than that predicted in Byram's fireline intensity equation ( Figure 3 1). Fire rate of spread (R OS) was faster in the drier (1.17 mmin 1 ) versus the wetter (0.61 mmin 1 ) FMC treatments (P=0.007), but was not affected by fuel loading treatments (P=0.446). Fuel consumption was high across all burns, ranging from 84 to 99%, but did not differ across FMC (P=0.130) or fuel loading treatments (P=0.387) (Table 3 1). Maximum surface temperatures differed across fuel loading treatments (P=<0.001), but not across FMC (P=0.887). Temperatures a t the fuelbed surface 1 ), but 1 fuelbeds, respectively ( Figure 3 fuelbed surface also differed across fuel load (P=0.002), but not between the low and
64 moderate FMC treatments (P=0.547). Lethal heating occurred for long durations and increased with fuel loading, with 9.480.73 min of lethal heating d uring burning in low fuel loads (10 Mgha 1 ) and 14.251.14 and 19.930.91 min during burning of the moderate (20 Mgha 1 ) and high fuel loads (30 Mgha 1 ), respectively ( Figure 3 3). Maximum belowground temperatures differed across the 2, 5, and 8cm s oil depths, heating was influenced both by fuel moisture ( Figure 4, P<0.001) and fuel loading ( Figure 5, P<0.001). The dr y FMC treatment resulted in greater soil heating com pared with the wet FMC treatment and higher fuel loading resulted in greater soil heating (P<0.001), but differences were only detected between the lowest (10 Mg/ha) and the two higher fuel loads (20 and 30 Mgha 1 ) using the Tukey Kramer post hoc comparis on ( Figure 3 5). The highest fuel load (30 Mg/ha) did not result in greater soil temperatures compared with the moderate fuel load (20 Mgha 1 ). Initial soil fuel loading (P=0.651), but was a significant covariate (P=0.006) in the general linear model. No interaction effects between soil depth, FMC, or fuel loading on soil heating were found to be significant using a GLM analysis of variance. Discussion Mechanical based fuels treatmen ts are being increasingly us ed by land managers to mitigate fire hazard and restore long unburned ecosystems. The ability to p redict potential fire behavior and subsequent fire effects is key to evaluating the effectiveness of these treatments. Heterogeneity of understory shrub and small tree biomass across space as well as time since disturbance will likely result in heterogenei ty of fuel loading on the forest floor following mowing treatments. Experimental manipulation of fuels
65 allows researchers to address how this heterogeneity influences potential fire behavior and ecological consequences. Here, we have determined the effect that varying fuel loading has on observed fire behavior characteristics as well as on measured aboveground and belowground heating from the burning of masticated palmetto/gallberry dominated pine flatwoods of the southeastern US. This work provides insigh t into masticated fuelbed fire behavior, and presents results relevant to assessing post mowing burning effects on soil ecology and residual vegetation. Although determining the effect of fuel loading across a larger range of moisture conditions would be informative, it was difficult under environmental constraints to do so. Environmental factors such as air temperature, relative humidity, and wind made it difficult to manipulate FMC to a large degree. However, the environmental and FMC conditio ns in this study would commonly occur during wildfire season in the region thus providing an indication of the effectiveness of such treatments at mitigating fire hazard in flatwoods forests. In testing our hypotheses, flame length and fireline intensity both increased with greater fuel loading and under drier FMC treatments, but rate of spread (ROS) only differed between FMC treatments, and consumption did not differ across any treatment. In support of our expectations, maximum surface temperatures incr eased with fuel loads, but did not differ between FMC treatments. Also as expected, maximum soil temperatures increased with greater fuel loads and under the dry FMC treatments, and surface increased with fuel load, but did not differ between FMC. Importantly, soil
66 Treatment Effects on F i re B ehavior and Above and Belowground T emperatur es Fuel loading and fuel moisture content (FMC) both independently increased flame length and fireline intensity. Even though the disparit y of the two FMC treatments was not substantial (~4%), the effect on fire intensity was significant, linked to a near doubling of ROS in the dr y treatment. Since neither ROS nor consumption differed across fuel loading treatments, fireline intensity was increased primarily from the increase in fuel biomass. Although higher fuel loading treatments were also greater in fu elbed bulk density, (to mimic that found in the field), the increased compactness of the fuelbed did not restrict the horizontal propagation of fire. The 10, 20 and 30 Mgha 1 fuel loading treatments resulted in 6, 9 and 12cm fuel depths and 16.7, 22.2, an d 25.0 kgm 3 fuelbed bulk densities, respectively. This range of fuel loading likely captures most understory and forest floor fuel loadings in pine flatwoods forests with a dominant palmetto and gallberry understory that have gone unburned (McNab et al. 1978) and where mowing treatments are most likely to be implemented. Maximum temperatures reached both at the fuelbed surface and belowground were influenced by fuel loading during the burning of these fuelbeds, yet fuel moisture content only influenced temperatures belowground. While maximum surface temperatures are likely reached instantaneously as the flaming front passes a given point, soil provides an insulation layer in which belowground heating likely depends both on the intensity of energy outpu t and the duration that heat energy is being transferred beneath the soil surface at a given location (Neary et al. 2005). The burning of drier fuelbeds resulted in greater soil heating and although ROS was faster, flame lengths were greater, compared wit h the wetter fuelbeds. Also, the wetting of fuelbeds for the
67 moderate FMC treatments did not increase soil moisture (P=0.847), averaging 9.91.1%, which may have otherwise subdued soil heating. Nonetheless, while maximum temperatures exceed The insulation capacity of the coarse soils on which these fuelbeds were burned may help mitigate the potential for lethal root heating du ring burning in these compact fuels. But high temperatures and long duration heating at the fuelbed surface could cause basal cambial damage to overstory trees. The duration of temperatures exceed ing es for each 10 Mgha 1 increase in fuel load. Although ROS differed across FMC but not fuel loading, duration of lethal heating differed across fuel loading, but not FMC. Lethal heating was not exclusively a function of flame residence time or fireline i ntensity, but was likely influenced by their combination, along with residual combustion following the passage of the flame front. Although total consumption did not differ across fuel load, the intensity and duration of residual combustion was likely gre ater in the heavier and more densely packed fuelbeds. The consequences of burning masticated fuelbeds are more likely to include damage to residual trees in long unburned flatwoods forests where fuel loads are high (Varner et al. 2005). Saw P almetto /G al l berry and Other Fuelbed Types C ompared While these fuelbeds were primarily composed of saw palmetto litter and some 1 h woody fuels, they were highly compact compared to that of typical pre mowing fuel strata in pine flatwoods (McNab et al. 1978). While f lame lengths observed here were not unlike those of other controlled experiments where compact masticated fuelbeds
68 from western US shrub fuels were burned (Busse et al. 2005, Kreye et al. 2011), maximum surface temperatures were somewhat lower and soil tem peratures were much lower (Table 2). Busse et al. (2005) developed an empirical model to predict maximum soil temperatures from fuelbed depth, soil moisture, and soil depth that drastically overestimates soil heating in our fuelbeds, ranging from 43 to 31 soil surface in this study. Although fuel depths across all three studies are comparable, fuel loading in the other two studies were substantially greater than ours (Table 2). Higher woody fuel loading in these other studies likely con tribute to the higher surface temperatures and much higher soil temperatures during burning due to higher total energy released per unit area, along with longer combustion times. Alth ough ROS was not measured in the above mentioned experiments, flaming times were observed to be much longer than those in this study. Busse et al. (2005) observed flaming times between 20 and 27 min in small fuelbeds (0.9 x 0.9m) and Kreye et al. (2011) observed 13 to 22 min of flaming from burning even smaller fuelbeds (38 x 26cm). Average flaming times in our study were 7 0.8 and 14 1.4 min in the low and moderate FMC treatments, respectively, over much larger (4 m diameter) fuelbeds. The greater foliar fuel component in the palmetto gallberry fuelbed is likely responsible for these differences. Fireline Intensity An exponential model fit t he relationship between fireline intensity and flame length from our study (R 2 =0.81), but this relationship differs from that of Byram's fireline intensity equation, which is commonly used in fire behavior/ fire effects prediction
69 software such as Behave Plus (Andrews et. al. 2005). The rate of energy output at the fire front is greater for a given flame lengt h than that observed by Byram (1959). Residual energy release following the passage of the flaming front may account for the different relationship observed in our fuelbeds compared with that observed by Byram (1959). Residual combustion following fronta l passage of the flaming front was anecdotally observed in this study, but to what extent it may account for the disparity between our observations and that of Byram (1959) is unknown. Nelson and Adkins (1986) also observed higher fireline intensities acr oss a range of flame lengths (46 144 cm) compared to Byram (1959) during both laboratory and field burning of slash pine ( Pinus ellio t tii ) needle litter beds with standing live saw palmetto. However, they found that flame lengths were relatively constant over a range of fireline intensities (98 370 kWm 1 ) when needle litter was burned without palmetto. Catchpole et al. (1993), on the other hand, found that fireline intensity increased with flame lengths according to Byram's (1959) equation when burning e ither excelsior or 6.35 mm sticks alone, but that flame length did not increase with fireline intensity when burning both excelsior and sticks in a mixed fuelbed. The higher fireline intensity observed in this study compared with Byram (1959) may indicate residual combustion during the burning of these fuels or there may be inconsistency in the relationship of fireline intensity and flame lengths across various types of fuelbeds. Nonetheless, any long term residual combustion during the burning of mastica ted fuelbeds may ultimately prolong heating at the forest floor and result in unintended ecological consequences such as tree mortality. The use of Byram's fireline intensity equation may therefore be inappropriate for estimating
70 fireline intensity from o bserved flame lengths in masticated saw palmetto gallberry fuels, especially if predicting possible fire effects on residual vegetation is of interest. The results of this study indicate that variation in fuel loading influences fire behavior and lethal he ating within masticated pine flatwoods fuelbeds. This variation will be important to managers to understand the effectiveness of these treatments to achieve management objectives, which often include retention of overstory trees. Mechanical fuels treatmen ts will likely occur where either flatwoods have not burned for several years, or in the wildland urban interface where the use of prescribed fire as a primary management tool is restricted. Mowing converts standing live fuels into compact surface fuels, so pre treatment standing biomass should translate into post treatment surface biomass. The ability to predict post mowing fire behavior and potential ecological effects enhan e mowing treatments. Further research is needed to e xplore how other fuels, moisture, and weather conditions affect fire behavior and effects in masticated fuelbeds. Such work would inform the development of additional fuel models to aid in fire prediction following mowing The majority of existing work has been conducted in compact masticated fuelbeds with low fuelbed depths, but where woody material is the primary fuel component. The gallberry/palmetto pine flatwoods of the southeastern US coastal plain is a widespread forest ecosystem, but with unique fuels compared to other North American fuel complexes. And while singular fuel models are already used to predict fire behavior in untreated flatwoods, this work suggests that the masticated gallberry/palmetto fuel complex also deserves a unique fuel mod el.
71 Table 3 1 Fire behavior characteristics from experimental burning of masticated understory vegetation of southeastern pine flatwoods across fuel loading and fuel moisture content treatments. Marginal and cell means are listed along with p values fr om GLM ANOVA. Flame Length Rate of Spread Consumption Fireline Intensity (cm) (m min 1 ) (%) (kJm 1 s 1 ) mean (SE) P mean (SE) P mean (SE) P mean (SE) P F MC a Low 111 ( 14 ) 0.001 1.17 (0.12) 0.007 93.6 (1.9) 0.130 593(1 16) 0.029 Moderate 67 ( 14 ) 0.61 (0.09) 97.0 (0.7) 317 (83) Fuel Load b ( Mgha 1 ) 10 49 ( 10 ) A <0.001 0.75 (0.19) 0.446 94.2 (2.2) 0.387 183 (47) A 0.003 20 91 ( 10 ) B 0.98 ( 0.16 ) 94.2 ( 2.2 ) 487 (81) AB 30 140 (14) C 1.00 (0.19) 97.6 (0.7) 773 (149) B FMC *Fuel Load Low/10 69 (7 ) 0.808 1.08 ( 0.27 ) 0.8 51 90.7 ( 3.5 ) 0. 343 260 (71) 0.758 Low/20 106 (16 ) 1.26 (0.13 ) 92.0 ( 4.0 ) 611 (84) Low/30 159 (6 ) 1.18 ( 0.27 ) 98.0 ( 0.6 ) 908 (207) Mod/10 29 (5 ) 0.41 ( 0.03 ) 97.7 ( 0.3 ) 105 ( 8) Mod/20 76 (7) 0.71 (0.19) 96.3 (1.8) 362 (101) Mod/30 114 (25) 0.75 (0.20) 97.0 (2.0) 569 (161) a Fuel moisture content treatment: low (8.9 0.6%) and moderate (12.92.0%) b Where fuel load was significant, similar letters within columns indicate no difference amongst means from the Tukey Kramer post hoc comparison ( =0.05)
72 Table 3 2 A comparison of observations from this study conducted in constructed fuelbeds of masticated palmetto/gallberry of southeaster n USA pine flatwoods and tha t of two other studies where burning experiments were conducted with constructed fuelbeds from masticated understory shrub vegetation of western USA forests. Study Dominant vegetation masticated Fuel moisture (%) Soil moisture (% VMC) Fuel depth (cm) Fuel load (Mgha 1 ) Flame l ength (cm) Surface t emperature Soil b temperature this study saw palmetto ( Serenoa repens) gallberry ( Ilex glabra ) 9 13 10 6.0 10 49 274 36 43 9.0 20 91 429 45 47 12.0 30 140 503 43 51 Busse et al. 2005 whiteleaf manzanita ( Arctostaphylos viscida ) common manzan ita ( A. manzanita ) 2 4 2.5 34 40 450 600 a 80 120 7.5 101 100 500 a 200 300 12.5 169 130 450 600 a 275 350 16 25 2.5 34 30 80 440 a 40 80 7.5 101 110 275 450 a 90 12.5 169 170 350 500 a 100 130 Kreye et al. 2011 common manzanita ( A. ma nzanita ) snowbrush ( Ceanothus velutinus ) 2.5 11.0 <5 7.0 69 95 436 771 131 208 a Range of peak temperatures were estimated, to the nearest 10, from temperature profile graphs in Busse et al.'s (2005) paper. b Soil temperatures recorded at 2.0cm (this st udy), 2.5cm (Busse et al. 2005), and 5 cm (Kreye et al. 2011 soil data unpublished) soil depths.
73 Fig ure 3 1 The relationship 1 between fireline intensity (kJm 1 s 1 ) and flame length (m) during the burning of fuelbeds created from masticated palmetto/ gallberry dominated pine flatwoods understory (solid line, R 2 =0.81), compared with Byram's (1959) fireline intensity equation 2 (dotted line).
74 Fig ure 3 2 The effect of fuel loading on maximum temperatures reached at the fuelbed surface during the burnin g of fuelbeds created from masticated palmetto/gallberry dominated pine flatwoods understory. Temperatures differed amongst all three fuel loading treatments using the Tukey Kramer post hoc comparison of the means.
75 Figure 3 3 The effect of fuel loadin g on the dur ation of aboveground surface heating created from masticated palmetto/gallberry dominated pine flatwoods understory. Lethal heating differed amongst the three fuel loading treatment s using the Tukey Kramer post hoc comparison of the means.
76 Fig ure 3 4 The effect of fuel moisture content (FMC) on soil heating (maximum temperatures) at three soil depths during the burning of fuelbeds created from masticated palmetto/gallberry domin ated pine flatwoods understory. note: FMC: low (8.90.6%) and moderate (12.92.0%).
77 Figure 3 5 The effect of fuel loading on soil heating at three soil depths during the burning of fuelbeds created from masticated palmetto/gallberry dominated pine f latwoods understory.
78 CHAPTER 4 FIRE BEHAVIOR AND EF FECTS IN MASTICATED PINE FLATWOODS ECOSYSTEMS OF FLORID A, USA Background The use of mechanical fuels treatments to reduce fire hazard in forest and shrub ecosystems has become a common management prac tice, however there are few empirical studies to elucidate the effectiveness of such treatments by quantifying fire behavior following their implementation. While treatments may be used as a stand alone option, they are often used as a pre treatment strat egy to reduce fire hazard during follow up prescribed burning. These types of treatments are being widely implemented across the United States (Glitzenstein et al. 2006, Kane et al. 2009, Kobziar et al. 2009, Brockway et al. 2010, Menges and Gordon 2010) and elsewhere (Molina et al. 2009, Castro et al. 2010), ranging in scale from a few to several thousand hectares. In addition to reducing fire hazard, fuels treatments are often conducted to restore long unburned ecosystems with goals of retaining mature o verstory trees and enhancing resistance to future fire (Agee and Skinner 2005). Evaluating the effectiveness of mastication type fuels treatments at reducing fire behavior and overstory resistance to post treatment burning is vital to determine treatment effectiveness. Mastication is a mechanical treatment that a lters fuel structure through mowing, shredding, or chipping understory shrubs and small trees. Front end or boom mounted equipment, attached to ground based equipment, is used to manipulate unders tory fuels with little impact to ground fuels or soils. Horizontal and vertical fuel continuity is disrupted, however total fuel loading is not reduced (Kane et al. 2009, Kobziar et al. 2009, others). Following treatment, masticated debris is either left on site or burned as
79 a follow up treatment to reduce surface fuel loading. While research is being conducted to evaluate potential biomass utilization, prescribed burning will likely remain a feasible option to remove masticated surface fuels following t reatments. Current work evaluating fire behavior and effects in masticated fuels is limited and much of it has been focused on western US ecosystems (Busse et al. 2005, Bradley et al. 2006, Kobziar et al. 2009, Knapp et al. 2011, Kreye et al. 2011). Furth ermore, while studies have used various approaches to address problems at different scales, empirical studies determining treatment effects on fire behavior at the stand scale are still few. Stand scale research in a variety of ecosystems will be needed t o not only evaluate the effectiveness of mastication at altering fire behavior, but also to determine whether results of fire behavior studies at smaller scales translate to scales in which treatments are being implemented. Small scale laboratory experimen ts have elucidated some understanding of the effects of particle or fuelbed scale properties on moisture dynamics (Kreye et al. 2012), fire behavior and potential fire effects (Busse et al. 2005, Kreye et al. 2011) in masticated fuels. While these studi es quantify the influence of fuelbed level properties on fire related metrics, it is unclear if such influences translate to a larger scale. Also, field studies at the larger scale have varied in results regarding the effectiveness of treatments at mitiga ting fire hazard (Bradley et al. 2006, Glitzenstein et al. 2006, Kobziar et al. 2009, Knapp et al. 2011). Treatment effectiveness will likely vary across ecosystems due to differences in pre treatment fuel structure. Mastication is being conducted in shr ub ecosystems with no overstory, in forest ecosystems with dense understory trees, as well as forest ecosystems that are dominated by shrub
80 understories. More empirical studies that quantify fire behavior and effects in masticated sites across several eco systems will enhance our understanding of fuel treatment effectiveness, but will also increase our general understanding of fire behavior and effects across masticated fuels that vary in fuelbed structure and properties. While mastication in forest and s hrub ecosystems often results in compact woody dominated fuelbeds (Kane et al. 2009, Battaglia 2010), mastication ("mowing") in pine flatwoods dominated by palmetto and gallberry shrubs in the understory results in fuelbeds comprised mostly of foliar litte r with a smaller percentage, by mass, of sm all diameter woody fuels ( Ch 2). Although compact, such fuelbeds will likely result in fire behavior and effects that are unique in comparison to those studied elsewhere. While small scale burning experiments ha ve revealed precise control of surface fuel loading over fire behavior in masticated debris collected from treatments in this ecosystem (Ch 3 ), it is unclear if surface fuels will control fire behavior at the stand scale in an ecosystem where shrubs recove ry quickly (Ch 2 and Ch 5). The objectives of this study were to 1) determine the effectiveness of mowing at reducing fire behavior at the stand scale in an ecosystem where masticated residues are primarily litter dominated and where shrub recovery is rapi d; 2) determine if surface fuels or shrub fuels controlled fire behavior six months following mowing ; 3) determine if fire induced tree mortality would increase as a result of burning in masticated treatments; and 4) evaluate the accuracy of current model s in predicting fire behavior following mowing
81 Methods Mechanical fuels treatments were conducted in the Osceola National Forest (ONF) in northern Florida, USA in pine flatwoods communities that had gone unburned for several years and wh ere fuel accumulations posed a hazard within the wildland urban interface (WUI). Pine flatwoods on the ONF are dominated by slash pine ( Pinus elliottii var. elliottii (Engelm.) and/or longleaf pine ( Pinus palustris Mill.) in the overstory and by saw palme tto ( Serenoa repens (Bartr.) Small) and gallberry ( Ilex glabra L. (Gray) shrubs in the understory. Mechanical mastication ("mowing") was used to reduce the height of understory fuels for re introduction of prescribed fire, and to reduce fire hazard in are as abutting communities, highways, and private pine plantations. Treatments occurred in mature pine flatwoods (ca. 80 yrs old) lacking a mid story and where the primary fuel strata altered during mowing was understory shrubs, including palmetto. Field E xperimental Burns For this study, two treatment locations were used to evaluate the influence of mowing on subsequent fire behavior and effects. The first location occurred within a 100 m wide buffer masticated in August 2009 under ONF management plans, a nd burned in July 2010 for this study. The second location occurred in a block experimental design set up for long term evaluation of the effects of mowing and mowing in conjunction with prescribed burning on ecological attributes (Ch 5). The experimental block treatments included mowing (mow: M), mowing followed by burning (mow+burn: M+B), burn without prior mowing treatment (burn only: B), and no treatment (control: C). Treatments blocks were ap proximately 2 ha in size ( Figure 4 1) and replicated three times. One replication was burned and used for ecological assessment (Ch 5), however during burning operations two of three control plots were burned when
82 fire escaped from a burn only block and fire behavior could not be assessed during operational contr ol of the escaped fire. Therefore, the third replicate was not used in this study for fire behavior analysis, but was used to assess post fire tree mortality. M and M+B treatments were masticated in August 2010, just following burning in the buffer locat ion, and subsequently burned in February 2011. While the buffer treatments were located according to the management plan of the ONF, the experimental treatment block locations were selected for this study in sites with similar ecological attributes as tho se of the buffer units evaluated in this study. Therefore, we were able to evaluate the effects of mowing on subsequent fire behavior and fire effects during dormant season (winter) burning, typical of the management regime, using the experimental block t reatments, but also to compare fire behavior and effects between a dormant (winter) season burn and a growing (summer) season burn in masticated treatments. Nine plots were allocated to each burn treatments (winter B, winter M+B, summer M+B) s ystematically to better facilitate coordination between ignition operations and fire behavior observations at each plot during burning ( Figure 4 2 ) In the experimental blocks, three plots, per treatment, were allocated to each block. As mentioned above, only six plots pre treatment were used to asse s s burning behavior, however all nine plots per treatment were used to analyze tree mortality. In the buffer treatment, nine plots were allocated to two locations within the buffer. Of the nine, six were mon itored during burning and all nine were evaluated for tree mortality. Burning operations were conducted by the ONF fire management staff using strip head firing technique s ( Figure 4 1). Ignition patterns were directed such that strip head fires were loca ted far enough downwind of plot locations whereby a strip head fire ignited 15 20 m
83 upwind of each plot location would burn through plots prior to downwind backing fires nearing plot locations. Observations of upwind strip head fires burning through plot locations were such that a steady state ROS and intensity appeared to be reached prior to plot ignition. Because the timing of burns were constrained by weather and ONF resource availability, full vegetation and fuels measurements could not be conducted immediately prior to burning. Comprehensive sampling of vegetation and fuels were conducted four months prior to each burn within all plots, including those not used in fire behavior analysis, but used for tree mortality. Sampling occurred in March 2010 prior to the July 28, 2010 summer burns, and in October 2010 prior to the February 23, 2011 winter burns. Overstory, shrubs, and surface fuels were quantified within each plot (full sampling techniques described in Ch 2) during sampling ( Figure 4 2). Tr ee height, diameter at breast height (DBH), and crown base height (CBH) were measured for all trees and tree density, basal area, and q uadratic mean diameter (QMD) were assessed for each plot. Total shrub biomass and shrub foliar biomass was estimated fro m shrub measurements in two 14 m belt transects. Litter depth, duff depth, as well as litter, duff, and woody (1h, 10, and 100h) fuel biomass was estimated using four 10m fuel transects. It was assumed that surface fuels and overstory trees were relativ ely unaltered during the four months between full sampling and burning, however shrubs grow relatively quickly in this ecosystem so a quick assessment of shrub characteristics was conducted for each fire behavior monitoring plot on the day of burning. Whi le the same sampling techniques were used four months prior to burning in both locations, sampling on the day of the burns differed between the summer burn in the buffer and
84 the winter burn in the experimental plots. Average shrub height was measured in e ach plot in both locations on the day of the burn, however shrub cover was also estimated during burning in the experimental block treatments. During summer burning in the mowed buffer, three subsamples of surface litter were collected and pooled, at each plot, to determine surface fuel moisture, on a gravimetric basis. During winter burning in the experimental treatment blocks (B and M+B), fuel moisture samples were taken of surface litter, as in the buffer, but also of live shrubs to compare between mast icated and non masticated sites. Temperature, relative humidity, and wind speed were recorded hourly during all burns and the Keetch Byram Drought Index (KBDI) reported. KBDI is a indication of soil moisture conditions, and thus used as a coarse assessme nt of fuel conditions, and is reported on a county scale. While duff moisture was not estimated during this study, KBDI is an indication of relative duff moisture differences between summer and winter burning conditions. Fire behavior was estimated duri ng burning using plot level measurements. At each plot, rebar was used as a measurement device to estimate ROS and flame height. In each plot, three rebar were located 8 m apart and oriented in line with predicted wind direction and perpendicular to the anticipated flame front. Each rebar was exposed 3.0 m above the surface litter and marked in 50 cm increments using fluorescent paint. In the experimental treatment blocks (winter burn treatments), litter and duff pins were also placed at each rebar loca tion prior to burning. Litter and duff pins were put in the ground with the top flush with the litter or duff, respectively to determine pre burn and post burn surface fuel loading and to calculate consumption. Observers followed the flaming front throug h each plot, marked the time of the arrival of the flame front at each
85 rebar location, and estimated flame height as the flaming front passed each rebar. ROS was calculated as the time between the arrivals of the flaming front at successive rebar location s divided by their respective distance. Flame heights were averaged across all three rebar by plot. Within a week of burning, tree damage was assessed for all trees within each 200 m 2 circular plot, including those not monitored during burning. Bole c har was measured in two ways: percent of the bole circumference at DBH charred and maximum char height. Crown damage was assessed by estimating crown volume scorched (CVS). CVS was visually estimated as the proportion of the crown volume that was scorch ed. Scorch occurs when foliage is desiccated from heating. Foliage is initially retained on branches and is typically reddish in color. Quick assessment was conducted to quantify CVS prior to needle loss. Scorched needles were not observed on the fores t floor at the time of damage assessments. Tree mortality was assessed one year following burning. To assess surface fuel consumption in the experimental treatment blocks, litter and duff depth consumed was measured at each litter and duff pin location. Depth consumed was measured as the distance between the top of pins and the fuel surface. Pre burn depth was measured as the distance between the top of the pin and the bottom of the respective fuel layer. Duff pins were located approximately 5 cm from litter pins so that pre burn litter depth could be estimated as the difference between the two if all litter was consumed. The percentage consumed was calculated from pre and post burn measurements. Litter mass consumed in the B treatments was estimated using bulk 3 Behm et al. 2004), while Litter mass consumed in M+B treatments, as well as duff mass
86 consumed for both B and M+B treatments, were estimated using litter bulk den sity 3 3 ) values determined in a similar masticated site nearby (Ch 2). Within a month of burning, all sites were fully re sampled for vegetation and surface fuels using the above sampling methods conducte d four months prior to burning. Because litter and duff pins were not used during the burning of the buffer locations (summer M+B), the pre burn and post burn surface fuel measurements, using the planer intercept method, was used to assess consumption in these locations. Data Analysis For fire behavior evaluation, comparisons were not made simultaneously across all three treatments (winter B, winter M+B, and summer M+B), rather planned comparisons were made between winter B and winter M+B treatments, and then between the winter M+B sites and the summer M+B sites. Therefore, separate analyses isolated the effect of mowing on fire behavior by comparing B and M+B sites burned in adjacent experimental treatments and on the same day, but also evaluated the ef fect of season of burn between two mowed treatments. For each analysis, pre burn vegetation and fuels measurements were compared between respective treatments. Pre burn vegetation and fuels were compared between the summer M+B and winter M+B treatments t o isolate seasonal effects on fire behavior and fuel consumption. Because measurements taken on the day of the burn differed between summer and winter burns, the comparison between summer M+B and winter M+B treatments were conducted using sampling techniq ues consistent between compared treatments. While litter and duff measurements were compared between the winter B and winter M+B treatments using measurements from litter and duff pins, respectively,
87 litter and duff measurements from fuel transects were u sed to compare consumption between winter M+B and summer M+B treatments. Therefore, errors associated with sampling methods would be consistent between treatments in respective comparisons. Fire behavior (ROS and flame height), fuel consumption (litter a nd duff), and overstory effects (CVS, char at DBH, and char height) were compared between burning treatments, again winter B versus winter M+B separately and winter M+B versus summer M+B separately. Plots used for these analyses were only those in which f ire behavior was assessed (6 per treatment). All subsamples, within plots, were averaged for analysis. Statistical comparisons were made using a two sample T Model assumptions were evaluated using the Shipiro Wilk test of normality and th e Modified Levene test for equal variance. Where model assumptions were not met, log or square root transformations were used to meet normality assumptions and the Aspin Welch Test used for unequal variances. To evaluate whether shrubs or surface litter was controlling fire behavior, linear regression was used to determine the correlation between flame height and shrub cover, shrub height, and litter mass. Linear regression was also used to determine the correlation between ROS and shrub cover, shrub he ight, and litter mass. Only data from the experimental winter burning blocks were used for this analysis. Shrub cover and shrub height were estimated on the day of the burn, as described above, and litter mass was determined from litter pins, as described above. To evaluate post burn tree mortality, three additional plots per treatment were included in analysis where fire behavior could not be monitored. Although fire behavior was not monitored in the extra plots, vegetation and fuels sampling was cond ucted in
88 the same manner as for those where fire behavior was monitored. Therefore, vegetation structure, fuel loading, fuel consumption, and tree damage characteristics were evaluated for all plots (9 per treatment). Mortality assessments between treatm ents (winter B, winter M+B, and summer M+B) were conducted with respect to measured tree characteristics (height and DBH) and tree damage characteristics (CVS,char DBH, char height). This allowed us to evaluate whether treatment differences in mortality w ere linked to treatment differences in damage or treatment effects were isolated from observed damage. While averages of CVS, char at DBH, and char height in the above analysis were compared between treatments, and using only the monitored plots during bu rning (6 per treatment), tree damage, across all treatment plots (9 per treatment), was evaluated at the individual tree level to evaluate mortality. Frequency distributions of CVS, char at DBH, and char height, were created for all trees separated by tre atment. Diameter distributions were created for each treatment, separately, with the number of trees dead, within diameter classes, indicated. Because a low number of trees died in this study, a rigorous statistical analysis of mortality rates could not be conducted. Evaluation of the effects of mowing or season of burning on tree mortality were assessed through simple evaluation of the number of trees dead within each treatment and determining whether tree size or tree damage observations (crown or pote ntial bole damage) were associated with mortalit y. P re burn vegetation and fuels as well as fuel consumption were also compared between burning treatments to determine if such differences could have attributed to post fire tree mortality. Since winter B and M+B treatments were established within an experimental design that included controls (C) and mow only (M) treatments (Chapter 5) with the
89 same level of replication and plot sampling techniques, background mortality was assessed in these C and M trea tments also. Modeled Versus Observed Fire Behavior Modeled fire behavior predications were compared to observed fire behavior in this field study. Rothermel's (1972) fire spread model was used to predict rate of spread 1 ), flame length (cm), and 1 ) at the plot level using fuel loading, fuel moisture, and weather conditions as measured above. Modeling was conducted in the BEHAVE PLUS Fire Modeling System (Andrews et al. 2008, version 4.0) and fuel parameters input from m easurements as a custom fuel model. Fuel moisture of 10h woody and live shrub foliage was not m easured in the summer M+B plots and 100h fuel moisture was not measured during any burning L ive fuel moisture was set at 100% a reasonable assumption based o n measured values in the winter burns, and a value recommended in BEHAVE PLUS when moisture is unknown (Andrews et al. 2008 ). 1 0h woody fuel s accounted for a small proportion of surface fuels compared to the finer fuels and their moisture content was much higher than finer fuels during the winter burns, thus they did not likely con tribute to flaming front combustion 100h fuels were rare. Also, 10h and 100h woody fuel consumption was difficult to measure in this study due to increases in these larger fue ls following burning likely due to input from tree damage above or exposure during surface combustion Post burning consumption values are used to estimate fireline intensity during burning. Therefore, only litter and 1h fuels were used as inputs into t he prediction model and their consumption used to calculate fireline intensity during observed burning. Fuelbed depth was input as average shrub height and live woody fuel loading as total shrub foliar biomass, the portion of shrubs involved in flaming co mbustion during burning.
90 Modeled outputs, by plot, were compared with the ROS, flame length, and fireline intensity observed during field burns. Flame lengths predicted in the model are measured from the top of the fuelbed to the flame tip along the flame axis, even if bent. Flames observed in the field were measured by vertical flame height and flame angle could not be observed from behind the flaming front. Observed flame length was nder the light wind conditions. Observed fireline intensity was calculated as fuel mass consumed 2 1 ). Fuel mass consumed included surface litter 1h woody fuels, and shrub foliage. Because nearly all shrub foliage was observed to have been consumed during burning, measured pre burn shrub foliar mass was assumed to have been consumed. Observed and predicted values were compared using linear regression. Results Winter M+B versus Winter B Treatments Regarding differences between winter B and M+B treatments, overstory vegetation did not differ between treatments, however shrub s were much reduced in M+B treatments co mpared to B treatments (Table 4 1). Tre e density, BA, QMD, height, and CBH averaged 336 tph, 17.1 m 2 per ha 25.9 cm, 20.9 m, and 14.9 m, respectively, across treatments. Shrub cover (p<0.001), shrub height (p<0.001), shrub biomass (p<0.001),and shrub foliar biomass (p<0.001) all were lower in M+B versus B treatments before burning In M+B treatments, shrub cover, height, total biomass, and 1 1 respectively. While in B treatments, cover, height, total biomass, and foliar biomass averaged 78 %, 145
91 1 and 1 respectively. Litter depth was lower in the M+B treatments (5.7 cm) compared with B treatments (7.6 cm ) (p=0.005, how ever litter mass, averaging 12.8 1 was higher than the 8.8 1 in the burn only sites (p=0. 002 ). Duff depth (p=0.1 16), averaging 3.8 cm, and duff mass (p=0.116), averaging 1 did not differ between treatments 1h woody fuel mass was higher in the 1 1 ) (p=0.015), but 1 d 1 also did not differ (p=0.534). Live fuel moisture did not differ between treatments (p=0.140), averaging 114 %, nor did 10h woody fuel moisture (p=0.465), averaging 24.4 %, however litter moistu re in M+B treatments (12.1%) was lower than in B treatments (17.8%) (p=0.047). KBDI during burning was 107. 1 While treatments could not be burned simultaneously to avoid any potential differences in weather conditions, treatments were burned from between 11:00 to 14:30 to avoid drastic differences in weather conditions. Flame heights during burning in M+B treatments ( 1.1 m) were one third of those observed in B treatments (3.3 m) (p=0.003), however ROS (3.4 1 ) di d not differ (p=0.150) (Table 4 2). Litter mass consumed was higher in M+B treatments ( 1 ) compared to B treatments (7.6 1 ) (p=0.026 ), however the proportion of l itter and duff consumption, averaging 85% and 2%,respectively, did not differ ( p=0.819, p=0.341) nor total duff mass consumed (0.6 1 ,p=0.341) ( p=0.997) Crown scorch averaged 45% across treatments and did not differ ( p=0.158). Bole char at DBH and maximum bole char height, however, were
92 marginally different (p=0.99 for both). Percent of bole circumference charred at DBH was 866% in M+B treatments and 972% in B treatments, and char height was 5.50.6 and 7.40.9 m i n M+B and B treatments, respectively. While shrub consumption was not quantified, nearly 100% of the understory area was burned and almost all shrub foliage consumed during burning ( Figure 4 3). Flame heights, across all treatment plots pooled, were cont rolled by both shrub cover (R 2 =0.80, p<0.001) and shrub height (R 2 =0.63, p=0.002), but not by litter mass (p=0.962) ( Figure 4 4). Shrub cover and height were also correlated (r=0.862). There was also some evidence that rate of spread was controlled by sh rub cover (R 2 =0.31, p=0.058) and shrub height (R 2 =0.27, p=0.084), but not by litter mass (R 2 =0.000, p=0.991). When conducting regression of fire behavior and fuel components within treatments, shrub cover was marginally related to flame heights in B sites (R 2 =0.575, p=0.081) and in M+B sites (R 2 =0.628, p=0.060), however shrub height was significantly related to flame heights in the B treatments (R 2 =0.712, p=0.035), but not in the M+B treatments (R 2 =0.003, p=0.914). Litter mass was not related to flame hei ghts in B (R 2 =0.063, p=0.631) or M+B (R 2 =0.070, p=0.613) treatments. There was slight evidence, however, that litter moisture influenced flame heights in M+B treatments (R 2 =0.548, p=0.092) but not in B treatments (R 2 =0.000, p=0.998). Live moisture was n ot a significant factor on flame heights in either treatments. Using multiple regression techniques revealed that only shrub cover was significantly related to flame heights and that all others were not significant with shrub cover in the model. ROS was not related to any quantified fuel characteristic within treatments.
93 Winter Versus Summer M+B Treatments In regard to winter M+B versus summer M+B treatments, Overstory conditions did not differ between treatments, however average tree height was slightly higher in the buffer M+B treatments burned in the summer (23.30.9 m) compared to the experimental M+B treatments burning in the winter (21.00.7), but differences w ere marginal (p=0.054) (Table 4 3). Tree density, BA, QMD, and CBH averaged 299 tph, 21 m 2 per ha 29.9 cm, and 12.3 m, respectively. Four months prior to burning, shrub height, averaging 64 cm, did not differ between treatments (p=0.467). While shrub cover was not quantified on the day of summer burning, total shrub biomass, averaging 0.75 1 did not differ between treatments (p=0.663), nor did shrub foliar biomass 1 ) (p=0.648), quantified four months prior to burning. Surface litter depth (5.5 cm) and duff depth (4.4 cm) did no t differ between treatments, nor did litter ( 12.2 1 1 ) mass. Woody fuels, however, differed in the smallest 1 in summer M+B 1 in the winter M+B treatments. 1 1 and were sparse. 1 respectively, during winter burning, temperature (31 (61 76%) were both higher during summer burning and wind speeds, while still mild, 1 (Table 4 3 ). Surface litter moisture was slightly higher during summer burning (14.71.1 %) compared to winter burning (12.10.6 %), h owever differences were marginal (p=0.064). While KBDI was 107 during winter burns, KBDI was 425 during summer burns, indicating drier soil conditions during the day of the summer burn. Quantified fire behavior did not differ between the
94 summer and wint e r burns (Table 4 4). Average flame heights were 1.50.1 m during summer burns and 1.10.3 m during winter burns but did not differ (p=0.267). Rate of spread (ROS) was 5 1 1 during winter burning and did not differ (p=0.276). Proportion of litter consumed was lower (p=0.014) during summer burns (487%) compared to winter burns (714%), and total litter mass consumed was 1 ) 1 ) burns. Duff consumption, however, was greater in average proportion consumed during summer burns (3211%) compared to winter burns (53%), but variation was high an d differences were marginal (p=0.067). 1 ) versus 1 ) burns, with high variation and marginal differences (p=0.098). Crown scorch, averaging 31%, did not differ bet ween treatments (p=0.406) and maximum char height, averaging 5.1 m, did not differ (p=0.319). Percent of bole circumference charred at DBH was 649% following summer burns and 866% during winter burns, but were only marginally different (p=0.069). The p roportion of understory area burned was almost 100% during summer burns as was observed during winter burns. Modeled versus Observed Using the Rothermel (1972) fire spread model, within the BEHAVE PLUS fire modeling system (Andrews et al. 2008, version 4 .0), there was much variation with observed and predicted values of fire behavior when modeling across all treatment burns at the plot level ( Figure 4 5). Observed flame lengths were generally over predicted in the mowed treatments, while under predicted in the non mowed (B) treatments ROS was over predicted across all plots burned, even in the non mowed
9 5 treatment however the relationship between observed and predicted ROS in the B plots appear s better Although there was variation in predictability, w ith more values over than under predicted, o bserved fireline intensity appeared to bette r fit predicted values than the other metrics. As observed with ROS, t here was a wider range of variation around the relationship between predict ed and observed firel ine intensity in the mowed sites compared to the un mowed (B) sites. Tree Mortality Across all treatment plots (winter B, winter M+B, and summer M+B), 165 trees were assessed for mortality, all of which were alive prior to burning. Of 65 trees assessed in winter B treatments, 2 trees we re dead one year later (Table 4 5, Figure 4 6) ; both were long leaf pines <20cm in DBH ( Figure s 4 6 and 4 7). Of 61 trees assessed in winter M+B treatments, all were alive one year following burning, however of 47 trees a ssessed following summer M+B burns, 7 we re dead one year later (Table 4 5, Figure 4 6). As a reference for background mortality, 1 out of 61 trees assessed in controls died during the study (longleaf pine, 9 cm DBH) and 0 out of 52 trees assessed in unburn ed mowed sites died. Of the 7 dead trees in the summer M+B treatment, 2 were small diameter hardwoods, while the others were all longleaf pines with only one being <20cm DBH ( Figure 4 7). Across all treatments, dead trees spanned the entire range of tree sizes across sites ( Figure 4 7), however only following summer burns did trees >20cm DBH and >20 m in height die Also, while all but two of the trees that died had almost 100% crown scorch, the two trees with <20% crown scorch that died following summer M+B burning were relatively large trees ( Figure 4 7). M ore trees in the winter B sites had >90% scorch compared to both wint er and summer M+B sites ( Figure s 4 6 and 4 8), however there were also more smaller trees ( <2 0 cm DBH ) in the B
96 treatments ( Figure 4 6). M ost of the small trees were 100% scorch ed in B treatments however there were still more larger trees (>20 cm DBH) in the B treatments with >90% scorch than in either the winter or summer M+B treatments ( Figure 4 8). Distribution of maximum char heights across trees were relatively similar between treatments, except that there were more trees in B treatments with char >8 m tha n in both M+B burn sites ( Figure 4 8). Percent of bole circumference charred at DBH was high across all treatments, howeve r M+B sites burned in the summer had l ess trees with >90% char ( Figure 4 8). Discussion While mechanical fuels treatments are being widely implemented to mitigate fire hazard, it is difficult to conduct field level experiments to gather empirical data e valuating their effectiveness. This study determined the effectiveness of understory mowing at reducing fire behavior in a common forest ecosystem of the southeastern US but also determined shrub control over fire behavior following these treatments, eva luated model predictability, and evidenced a seasonal effect on tree mortality. Recent research has begun to characterize the post mastication fuel environment in various ecosystems, however much of this research has been focused in the western US (Hood a nd Wu 2006, Kobziar et al. 2009, Kane et al. 2009, Battaglia 2010) Published reports have shown that s urface fuels result ing from mastication of shrubs and small trees in these ecosystems are primarily composed of wood y fuels (Kane et al. 2009, Battaglia 2010) L aboratory scale fire behavior studies have revealed that burning in these compact, woody dominated fuelbeds result in long duration surface (Kreye et al. 2011) and soil (Busse et al. 2005) heating S ome field studies have also
97 shown unexpected tr ee mortality following burning in these treatments (Bradley et al. 2006, Knapp et al. 2011). Mastication ("mowing") in palmetto/gallberry pine flatwoods of the southeast ern US results in litter dominated surface fuels (Ch 2) much different than other are as studied This work broadens our understanding of fire behavior in masticated forests and shrublands in general, and provides insight into their effectiveness in this region. Flame heights were reduced by two thirds following mowing in this ecosystem, however shrubs began controlling fire behavior as soon as six months following mowing Small scale fire behavior experiments conducted with collected surface material following mowing in these sites revealed precise control of litter biomass over fire be havior (Ch 3 ), which was not evidenced in this study. Shrubs were much reduced in the treated sites, however their quick recovery resulted in a shrub type fuel model (Scott and Burgen 2005) soon after treatment as evidence d by their control over fire beh avior. While mastication in many shrub and forest ecosystems may result in a surface fuel type for some time, mastication in areas where shrubs resprout vigorously will likely return to a shrub fuel type quickly and treatment efficacy on fire behavior may be short lived. Results here indicate that if follow up prescribed burning is conducted soon after mowing treatments are effective at reducing fire intensity but as early as six months following mowing shrub s will influence fire behavior. Although shru bs controlled flame heights in this study, combustion in the surface litter created from mowing may still be an important management concern. There was evidence here that burning in mowed sites resulted in less crown scorch, likely due to lower flame heig hts (Van Wagner 1973). More trees were 100% scorched in the winter
98 burn only treatments compared to both winter and summer burning in mowed sites, however more tree mortality was evidenced following summer burning in mowed sites and even two large trees with little crown scorch died. Litter consumption during winter burns was high, but little duff was consumed. During summer burns, less litter consumption was observed, but there was some evidence of greater duff consumption, even though there was a lot of variation. While litter consumption in the summer burns were lower, on average, long duration heating from litter combustion (Ch 3 ) may have been enough to ignite duff during summer burns where soil conditions were likely drier, as indicated by higher KBDI. While high flammability in these historically frequently burned ecosystems may alleviate long duration surfa ce and soil heating (Gagnon 2010 ), long duration heating in surface material following mastication (Busse et a l. 2005, Kreye et al. 2011, Ch 3 ) in conjunction with duff accumulation in long unburned sites (Varner et al. 2005) may result in fine root or bole damage (O'Brien et al. 2010 a ). While southern pines are capable of recovering from substantial crown damage (Waldrop and Van Lear 1984, J ohansen and Wade 1987), effects of fine root or bole damage may result in delayed mortality (O'Brien et al. 2010 a ), and trees that survived one year following burning in this study could potentially still die. Although few trees that we assessed for morta lity died in this study, there was evidence that summer burning resulted in greater tree death and that tree damage typically attributed to greater flame lengths (crown scorch and bole char) didn't explain differences in mortality across burning treatments Larger sized plot s or transects would have been needed to inc orporate more trees in to our study allowing a more thorough statistical analysis of tree mortality. The effectiveness of mastication in reducing fire behavior may be
99 important for restorati on of long unburned sites, but timing of prescribed burns should be taken into consideration regarding potential ecological effects. Since mastication treatments only alter fuel structure and do not reduce fuel loading, follow up burning objectives will li kely include the consumption of surface fuels created from treatments. While litter cons umption was quite high in these burns, the mulching effect of moisture retention in mastic ated fuelbeds ( Kreye et al. 2012 ) may mean that attaining desired fuel consum ption may be difficult under wetter surface conditions. Although KBDI was higher during summer burning in this study, indicating drier soil conditions, there was some evidence of higher litter moisture during summer burns. In contrast, shrub reduction fo llowing mowing in this ecosystem may result in drier su rface fuels ( Ch 5), potentially due to increased solar radiation or surface winds. Litter moistures, during winter burns, were lower in mowed treatments versus un mowed treatments. In long unburned f orests, where mowing is likely to occur, duff accumulation may be heavy and burning under drier conditions could result in high tree mortality (Varner et al. 2007). When prescribed burning is used as a management tool on a large scale, as in this region, meeting frequent fire cycles can be difficult under the constraints of "burn windows". If burning conditions required for surface fuel consumption are such that burn windows are narrowed, it may be difficult to burn masticated sites soon enough to avoid s ubstantial shrub recovery, but also during conditions to avoid potential tree mortality. Developing treatment regimes that incorporate mowing and burning may require strategic timing to meet management goals without resulting in unintended ecological cons equences. Not following up quickly
100 enough with fire may result in dense surface fuels on top of accumulated duff, but under heavy shrub loading within just a few years following treatments. Model predictability of fire behavior in these treatments varie d with regard to the fire metric. Flame lengths were over predicted in mowed sites and ROS over predicted for both mowed and un mowed sites. Flame lengths were adjusted for an assumed flame tilt of 30 under the light wind conditions. It is difficult, however, to quantify actual flame length during burning, especially in shrub fuels. From observations in the field, a 30 tilt is likely a liberal estimate, however the adjustment from vertical flame he ight is only an increase of 15%. One additional point is that flame heights measured during burning extended from the litter surface to the top of the flame, however model predictions are such that flame length extends from the top of the fuelbed surface, in this case being average shrub height. If estimated flame lengths from field burning were adjusted to include only the flaming portion above the average shrub height, model performance would be even poor er with drastic over estimations of flame length. Flame length above the forest floor is likely a more important metric as a tool to assess fire suppression tactics or controllability during prescribed burning. Fireline intensity a ppeared to be better predicted in un mowed sites, compared to the mowed sites. Fireline intensity is calculated from fuel consumption and ROS, and also should be related to flame lengths (Byram 1959). A tighter relationship between observed and predicted ROS in B plots likely attributed to the tighter relationship of observe d and predicted fireline intensity in B plots, however intensity was much closer to observed values. One major shortcoming of the Rothermel (1972) model is that is assumes a homogenous fuelbed. In these shrub fuels there is a vertically oriented shrub f uel layer
101 above a denser horizontally oriented surface layer even when un mowed The model is quite sensitive to fuelbed bulk density and the heterogeneous nature of these fuels is exacerbating following mowing where higher surface fuel loads are even m ore compact, but under a quickly recovering shrub layer. In chapter 3, fireline intensity in these masticated surface fuels was observed to be greater per unit flame length than Byram's (1959) relationship, which is used to predict flame lengths in the mo del. The relationship between flame lengths and fireline intensity may not be as clear in such fuel scenarios, especially in a heterogeneous fuelbed where shrubs are burning above combustion in a much denser surface layer beneath. Using current fire modeling techniques to assess fuel treatment effectiveness may be problematic (Varner and Keyes 2009) and their use likely depends upon the specific ecosystems in which treatments occur. Research that compares model predictions with fire behavior observed in masticated treatments is lacking and the few that exist var y in regard to how well model s predicted fire behavior (Kobziar et al. 2009, Knapp et al. 2011). There was much variation regarding the accuracy of model predictions and observations of fire b ehavior at the plot level in the study, however avera ge predictions across sites may be sufficient to predict fire behavior at the stand scale at least in regard to fireline intensity Models are generally used as a prediction tool across a site and are typically not used to predict fire behavior at a more localized plot scale. Inaccurate predictions at our plot scale does not necessarily mean that the underlying physical processes involved in combustion are not accurately portrayed in the model. Spatial variation in fuel structure likely attributes to spatial variation in fire behavior which may not be fully captured, even at our plot level, when average fuel
102 characteristics are input to the model Mismatches between observed and predicte d outputs may occur for several reasons. Whether mismatches occurred because the model has fundamental errors, if fuel or weather inputs were inaccurate at our scale, if model parameters are inappropriate for these kinds of fuelbeds, if the uncertainty of the model doe s incorporate the variation we observed, or if our methods of observation were inaccurate is unknown. But, even if models may be sufficient to predict fire intensity, their use as a predictor of fire effects may be limited where dense surface fuels bene ath a burning shrub layer, may be generating localized heat for long durations One of the model assumption is fuelbed h omogeneity, which does not accurately represent a dense masticated fuelbed beneath a shrub layer Fuel models developed specifically f or masticated sites will need to incorporate the heterogeneous aspect of the fuelbed to better predict potential heating of surface and soil layers, and ultimately fire effects to the ecosystem. Empirical evaluation of the efficacy of mechanical fuel treatment at altering fire behavior and effects is difficult, especially under experimental control. We have observed here that the mitigating effect of mowing on fire hazard in a common pine ecosystem of the southeastern US is applicable to observed fir e behavior, but not necessarily to fire effects. When planning treatment regimes that incorporate both mowing and prescribed fire, timing will likely be critical in order to mitigate rapid fuel recovery and burn under conditions to avoid potential unfores een consequences, all while meeting management objectives.
103 Table 4 1 Weather, overstory, and fuel conditions during experimental burning of masticated (mow+burn) and untreated (burn) stands of palmetto/gallberry pine flatwoods in northern Florida, USA. Burning Conditions Burn Date Temp RH Windspeed Litter Live 10h KBDI C % km hr 1 ----------% moisture -----------Mow+Burn 23 Feb 2011 17 24 47 62 1.6 4.8 12.1 (0.6) A 117 (3) A 20.9 (6.6) A 107 Burn Only 17.8 (2.4) B 110 (3) A 27.8 (5.6) A 107 Overstory Tree Density Basal Area QMD Height CBH t rees h a 1 m 2 ha 1 cm m m Mow+Burn 307 (6 4) A 18.9 (4.4) A 27.8 (1.6) A 21.0 (0.7) A 14.7 (0.9) A Burn Only 365 (63) A 15.2 (1.7) A 23.9 (1.9) A 20.7 (1.6) A 15.1 (1.1) A Understory Fuels Shrub Cover 1 Shrub Height 1 Shrubs Shrub Foliag e % cm -------------------Mg ha 1 ----------------Mow+Burn 32.5 (3.6) A 58 (13) A 0.6 (0.3) A 0.4 (0.2) A Burn Only 77.5 (4.0) B 145 (8) B 4.4 (0.5) B 4.1 (0.5) B Surface Fuels Litter Depth Duff Depth Litter Duff 1 h 10 h 100 h -----------cm -----------------------------------------Mg ha 1 ---------------------------Mow+Burn 5.7 (0.4) A 3.0 (0.5) A 12.8 (1.0) A 33.6 (5.5) A 1.1 ( 0.2) A 2.1 (0.3) A 1.1 (0.6) A Burn Only 7.6 (0.2) B 4.5 (0.7) A 8.8 (0.3 ) B 49.5 (7.4) A 0.5 (0.1) B 1.1 (0.4) A 0.7 (0.3) A Note: Values sharing letters within columns are not statistically differ ent (Tukey
104 Table 4 2 Fire behav ior and effects from burning of masticated (mow+burn) and unmasticated (burn only) palmetto/gallberry pine flatwoods. Fire Behavior Consumption Overstory Fire Effects Flame Ht ROS Litter Duff Litter Duff Scorch Char Char Height m m min 1 ------Mg ha 1 --------------% ------% % m Mow+Burn 1.1 (0.3) A 3.4 (1.0) A 10.6 (0.8) A 0 .0 (0 .0 ) A 83 (4) A 0 (0) A 37 (8) A 86 (6) A 5.5 (0.6) A Burn Only 3.3 (0.5) B 7.1 (2.1) A 7.6 (0.8 ) B 1.1 (1.1) A 86 (8) A 3 (3) A 53 (6) A 97 (2) A 7.4 (0.9) A Note: Valu es sharing letters within columns are not statistically different (Tukey indicates marginal differences (p<0.10) Marginal difference (p<0.100)
105 Table 4 3 Comparison of burning conditions (weather, overstory, and fuels) between a summer and winter burn in masticated palmetto/gallberry pine flatwoods of northern Florida, USA. Burning Conditions Burn Date Temp RH Windspeed Litter Moisture KBDI C % km hr 1 % S ummer 28 Jul 2010 31 34 61 76 1.6 7.2 14.7 (1.1) A 425 Winter 23 Feb 2011 23 24 47 49 1.6 2.7 12.1 (0.6) A 107 Overstory Tree Density Basal Area QMD Height CBH trees h a 1 m 2 ha 1 cm m m Summer 290 (27) A 23.1 (3.0) A 32.0 (2.6) A 23.3 (0.9) A 15.8 (0.8) A Winter 307 (64) A 18.9 (4.4) A 27.8 (1.6) A 21.0 (0.7) A 14.7 (0.9) A Unders tory Fuels Shrub Height 1 Shrubs Shrub Foliage cm -----------------------Mg ha 1 ----------------Summer 69 (7) A 0.9 (0.5) A 0.5 (0.2) A Winter 58 (13) A 0.6 (0.3) A 0.4 (0.2) A Surface Fuels Litter Depth Duff Depth Litter Duff 1 h 10 h 100 h -----------cm -----------------------------------------Mg ha 1 ---------------------------Summer 4.9 (0.7) A 5.3 (0.8) A 10.9 (1.6) A 58.8 (9.4 ) A 4.1 (1.0) A 6.6 (0.6) A 2.5 (1.1 ) A Winter 6.0 (0.4) A 3.5 (0.6) A 13.4 (0.9) A 38.8 (6.5) A 1.1 (0.2) B 2.1 (0.3) B 1.1 (0.6) A Note: Values sharing letters within columns are not statistically different (Tukey 05) M arginal differen ce (p<0.10)
106 Table 4 4 Fire behavior and effects between summer (July) and winter (Feb) burning of masticated palmetto/gallberry pine flatwoods. Fire Behavior Consumption Overstory Fire Effects Flame Ht ROS Li tter Duff Litter Duff Scorch Char Char Height m m min 1 -------Mg ha 1 --------------% -------% % m Summer 1.5 (0.1) A 5.9 (1.8) A 5.5 (1.3) A 23.1 (10.1) A 48 (7) A 32 (11) A 25 (11) a 64 (9) A 4.7 (0.6) A Winter 1.1 (0.3) A 3.4 (1.0) A 9.6 (0.9) B 2.6 (1.9) A 71 (4) B 5 (3) A 37 (8) a 86 (6) A 5.5 (0.6) A Note: Values sharing letters within columns are not statistically differ ent (Tukey M arginal differences (p<0.10)
107 Table 4 5 Number of tree s dead or alive across three treatments at one year following burning in palme tto/gallberry pine flatwoods. Burn Only a (winter) Mow+Burn b (winter) Mow+Burn c (summer) Total Dead 2 0 7 9 Alive 55 61 40 156 Total 57 61 47 165 Note: All trees were alive prior to burning. a Non masticated, burned Feb, 2011 b Masticated Aug, 2010, burned Feb, 2011 c Masticated Aug 2009, burned Jul, 2010
108 Figure 4 1 Experimental mowing and burning t reatments in pine flatwoods in northern Florida, USA (Osceola National Forest). Systematic plot locations are indicated. Burn only and mow+burn treatments burned with strip head firing techniques (white arrows indicate fire movement).
109 Figure 4 2 Ex ample of plot locations within the buffer treatments. Sampling within plots were the same for both buffer and experimental block treatments. All fuel transects were randomly oriented.
110 Figure 4 3 Fire behavior in experimental mowing and burning treatments in pine flatw oods of northern FL, USA. Burn only treatments were not masticated, mow+burn treatments were mastic ated 6 months prior to burning.
111 Figure 4 4 Fire behavior measurements (rate of spread, above; flame height, below) as a function of shrub cover (left), shrub height (middle), and litter mass (right) during the burning of mowed and un mowed experimental treatments in pine flatwoods.
112 Figure 4 5 Observed versus predicted fire behavior across burning treatments within mowed (M+B) and un mowed (B) p almetto/gallberry pine flatwoods burned in the winter (Feb) and mowed treatments burned in the summer (M+B summer). Solid line, 1:1 ratio; Dashed line, linear regression.
113 Figure 4 6 Crown scorch (%) versus tree diameter (DBH) (left) and t ree mortality within diameter distributions (right) across burn only (top) and mow+burn (middle) treatments burned in the winter (Feb) and mow+burn treatments burned in the summer (July) (bottom).
114 Figure 4 7 Tree mortality across individual tree c haracteristics (height and DBH) and tree damage (crown scorch and bole char height) following burning in masticated and non masticated treatments in palmetto/gallberry pine flatwoods. The height vs DBH graph indicates the only 2 hardwoods in the study (bo th died) and the only 2 trees that died in the burn only treatment, all other dead trees occurred in the masticated treatment burned in the summer. Trees 1 and 2 are indicated in both graphs and were both large trees with little crown scorch that died fol lowing summer burning following mowing
115 Figure 4 8 Distribution of crown scorch (top) bole char height (middle) and percent bole circumference charred at DBH (bottom) across burn only and mow+burn treatments burning in the winter (Feb) and mow +burn treatments burned in the summer (July)
116 CHAPTER 5 EFFECTS OF MECHANICA L FUEL TREATMENTS AN D PRESCRIBED BURNING ON VEGETATION, MICROCLI MATE, AND SOILS IN PINE FLATWOODS EC OSYSTEMS OF FLORIDA, USA Background Fire is a critical ecological process in many ecosystems worldwide. In many ecosystems, however, fire exclusion has resulted in increased accumulation of fuel, often leading to increased wildfire hazard. The use of prescribed burning as a management tool to reduce fuel loads is often difficult due to the increased fuel biomass and the challenges of the wildland urban interface, where human development is interspersed with wildlands. In addition, factors associated with climate change may likely result in increased risk of wildfire occurrence and the extent of areas burned (Westerling 200 6) The use of mechanical treatments to reduce fuel biomass or to alter fuel structure as a means to mitigate such hazards has become widespread. Treatments may also be used for restoration where fire sensiti ve species have invaded as a result of fire exclusion. Treatments are used as a stand alone management tool or in conjunction with prescribed burning to consume treatment residues or restore a managed fire regime to the ecosystem. The ecological consequen ces of these treatments are poorly understood. Mec hanical mastication has become widespread in recent decades, and is used to alter fuel structure in both forest and shrub dominated ecosystems (Hood and Wu 2006, Glitzenstein et al. 2006, Kane et al. 2009, Kobziar et al. 2006, Battaglia et al. 2010) Understory shrubs or small trees are masticated (chipped, mowed, etc.) via front end or boom mounted mastica tion heads attached to various types of mobile machinery (Gyrotracks, skidders, etc.). Immediate resu lts of such treatments include a reductio n in
117 fuel height and a compaction of the fuelbed, however no reduction in fuel loading occurs. Mastication's immediate effects consist only of the rearrangement of fuel structure(Kobziar et al. 2009, Kane et al. 20 09). Current research has focused on mastication of shrubs and small trees in western US ecosystems where the resulting fuelbed is primarily composed of small diameter woody material fractured through the mastication process (Kane et al. 2009, Kobziar et al. 2009, Battaglia et al. 2010, Kreye et al. 2011). Initial research in these fuels indicates that while fire intensity may be reduced immediately following treatments, compact fuelbeds may lead to long durations of burning (Kreye et al. 2011) with suffi cient heat to cause ecological change, such impacts to soils (Busse et al. 2005, Kobziar and Stephens 2006) or overstory trees ( Knapp et al. 2011). Long duration heating of fuelbed surfaces and underlying soils has been observed during laboratory burning in collected fuels from western sites (Busse et al. 2005, Kreye et al. 2011) as well as impacts to soil respiration in the field (Kobziar and Stephens 2006) H igher than expected fire intensity (Bradley et al. 2006, Kobziar 2009) and overstory mortality ( Knapp et al. 2011) have been observed from post mastication burning in field experiments as well as i ncreases in both native and non native understory species (Kane et al. 2010) A lthough laboratory experiments have indicated that compact woody fuelbeds following mastication enhances moisture retention (Kreye et al. 2012), field assessments of treatment impacts on moisture dynamics or micrometeorology have not been explored. Furthermore, impacts on soil nutrients and decomposition of surface litter have been given little attention in masticated sites, especially where sites are burned following treatment.
118 While some research on the ecological impacts of mastication have been conducted in the western US where resulting fuelbeds are primarily woody debri s (Kane et al. 2010 Knapp et al. 2011, Rhoades et al. 2012) mastication ("mowing") is being widely employed in palmetto/gallberry pine flatwoods of the southeastern US where surface fuels following treatments are largely composed of foliar litter and, to a lesser extent, very smal l diameter woody debris (Ch 2). In the sub tropical climate of the southeastern US, humidity and regrowth rates are higher, species composition and flammability differ, and consequences of fuels treatments to soil characteristic s are unknown when compared with the western US. Understanding these treatment effects is key to evaluating the usefulness of mechanical treatments in this region. Pine flatwoods represent the most widespread and prevalent forested ecosystem in the coasta l plain of the southeastern US Flatwoods occur on sandy soils of marine origin. In areas where seasonal inundation of water due to poor drainage occur, nutrient poor Spodosols are common, and understory shrubs, dominated by saw palmetto ( Serenoa repens (Bartr.) Small ) and gallberry ( Ilex glabra (L.) Gray ), occur under a canopy of longleaf ( Pinus palustris Mill.) or slash pine ( P. elliottii Engelm. ) with varying densities. While fires were historically frequent ( return interval <10 yrs) in these ecosyste ms, sites that have gone unburned for over ten years are being treated using mastication ("mowing") to reduce fire hazard on many public lands especially in the wildland urban interface Although soils are nutrient poor in these sites, understory vegetat ion resprouts vigorously following disturbance. The quick response of understory vegetation to disturbance make s this a unique ecosystem where the ecological effects of mastication have been inadequately addressed.
119 Existing work on mastication type fuel s treatment s in the southeast have focused on treatments in the more xeric scrub and sandhill ecosystems ( Brockway et al. 2009, Menges and Gordon 2010, ) These s tudies have shown that mechanical treatments alone are not as effective as burning for attain ing restoration goals that typically include reduction of woody plant cover and promotion of herbaceous species. I nitial reductions in tree density following mowing has been observed, but with a quick recovery of hardwoods and with substantial increases in understory plant cover but without desired gains in grasses and forbs (Brockway et al,. 2009). L ittle work exists in the primary literature regarding m astication in the wetter palmetto/gallberry dominated flatwoods ecosystems. The few studies addressin g ecological effects of understory fuels treatments in this ecosystem have focused on roller chopping (Schwilk et al. 2009) a treatment that differs from mastication in that a rolling drum, filled with water, is pulled across the ground resulting in great er soil disturbance than mastication (O'Brien et al. 2010 b ) Effects of roller chopping to understory plants have been found to be minimal in both flatwoods (Schwilk et al. 2009) and dry prairies (Watts and Tanner 2006), an ecosystem floristically similar to that of flatwoods, but without an overstory (Abrahamson and Hartnett 1990). When in combination with burning, however, these treatments may reduce shrubs and enhance herbaceous layers the common restoration goals, better than either roller chopping o r burning alone (Watts and Tanner 2006, Schwilk et al. 2009). Whether mastication treatments, where soil disturbance is less likely, would have similar effects in this commonly targeted ecosystem i s unknown. To evaluate ecological impacts from masticati on fuels treatments in pine flatwoods ecosystems, vegetation dynamics, microclimate, understory moisture dynamics, litter
120 decomposition, and soil nutrients were determined following mowing treatments in northern Florida, US The ob jectives of this study were to 1 ) determine changes in overstory, understory, and groundcover up to two years following mechanical mowing treatments in three s tand types of pine flatwoods; 2 ) measure decomposition rates of litter created by mechanical treatment; and 3 ) compare t he effects of mowing and m owing followed by prescribed burning on vegetation (overstory, understory, and ground cover), microclimate (air temperature and relative humidity), shrub and litter moisture, soil temperature and soil nutrients. Methods Mechani cal fuels treatments were conducted in the Osceola National Forest (ONF) in northern Florida, US in pine flatwoods communities that had gone unburned for several years and where fuel accumulations pose a hazard within the wildland urban interface (WUI). M esic p ine flatwoods on the O NF are dominated by slash pine and/or longleaf pine in the overstory a nd by saw palmetto and gallberry shrubs in the understory. Mature pine stands with moderate tree densit ies and open canop ied structure s are common but many younger pine plantations also exist Shrubs tend to dominate the understory, but grasses, mainly wiregrass ( Aristida spp.), may be common as well as other herbaceous plants. Climate is hot in the summer averaging 33 C, and mild, but variable, in the w inter with 1 7 to 19 C highs and lows as cold as 10 C (Chen and G erber 1991). Summer months are the wettest with precipitation occurring from frequent thunderstorms. Topography is flat and soils are primarily Spodosols of coarse textured marine deposits that are poorly drained. Mowing, a regional term for mastication, was used to reduce the height of understory fuels for re introduction of
121 prescribed fire, and to reduce fire hazard in areas abutting communities, highway s, and private pine plantations. Treatments occurred in mature pine flatwoods (ca. 80 yrs old) and a younger plantation (28 yrs old), both lacking a mid story and where the primary fuel strata altered during mowing was a continuous understory of saw palme tto and gallberry shrubs. Treatments used in this study occurred in two locations ( Figure 5 1), 1) a 100 m wid e and 6 km long buffer, mow ed in 2009, that included mature pine, mature pine that was recently burned (5 yr since fire), and young pine plant ation; and 2) three experimental blocks (8 ha ea) in mature pine that were m owed in summer 2010 and burned in spring 2011 to create the f ollowing treatments: mowing only (mow), m owing followed by burning (mow+burn), burn only (no m owing ), and control. Wit hin each of the three 8 ha experimental blocks, treatments were approximately 2 ha each ( Figure 5 1). Each block received each of the four treatments and treatments within blocks were systematically allocated to facilitate burning operations by the ONF ma nagement and to create an edge between each treatment and all other treatments. Soils in both treatment areas were sandy or sandy over loamy, siliceous, thermic, ultic aloquods (USDA Soil Survey). Mechanical treatments were conducted with forward mo unted mowing heads, with fixed cutters, attached to tracked ground equipment. The treatment prescription was that all understory shrubs and trees <20 cm DBH were to be mowed and all re sidue left on site. Mowing treatments differ from roller chopping trea tments, common in the region, in that mowing is not intended to disturb the soil. While roller chopping drums, filled with water, are pulled behind ground equipment with their full weight on the
122 ground, front mounted mowing heads are not fixed in position but are hydraulically controlled (up and down) by the operator. Burning treatments in the experimental blocks were conducted in February, 2011 by ONF personnel using strip head firing techniques Vegetation was ignited with hand held drip torches with a pproximately 20 m spacing between ignition lines and fire moving with the wind Conditions during burning were 17 24 C with relative humidity ranging from 47 62% and under light winds (1.6 4.8 km hr 1 ). Rate of fire spread was slow to moderate during burning (3.4 7.1 m min 1 ) and while flame heights averaged 1.1 m in mowed sites, they were higher (3.3 m) in the un mowe d sites (Kreye Ch 4). Nearly 100% of the area was burned with almost all shrub foliaged consumed and 85% of litter consumed, but with little to no duff (humus and fermentation layers) consumed. Vegetation Dynamics Mowing in the buffer treatment occurre d in August 2009. Vegetation was sampled immediately prior to treatment, and at 2, 8, 16, and 24 months following treatment. The 8 mos sampling period was conducted at the beginning of the growing season (Mar, 2010), 16 mos sampling after the growing sea son (Oct, 2010), and 24 mos in Aug, 2011. Pre treatment sampling plots were systematically located within the linear buffer and subsequently re sampled following treatment. Allocation of plots within stand types (mature N=12, mature/burned N=9, plantatio n N=6) were weighted based on the area represented by each stand type along the buffer. Plots were spatially arranged in triplets at 15, 45, and 75 m from the buffer edge, but arranged at a 45 angle between plots in reference the edge of the buffer ( Figure 5 2). They were spatially established by locating the ce nter of the stand type unit, to reduce edge influence from adjacent
123 stand types, and were arranged so that an equal number of plots w ere located on either side of the center of the unit. Within each plot, all trees 2.5 cm diameter at breast height (DBH, measured at 1.37 m above the ground) were measured for height and DBH within the entire 201 m 2 circular plot. Trees were measured before treatment, after treatment, and at two years following treatment. Basal area (m 2 ) and q uadratic mean diameter (QMD) were calculated for each plot, at each sampling period. Shrubs and tree saplings (<2.5 cm DBH) that were at least 0.5 m in height w ere tallied, by species, within two 1 8 m belt transects located at 4 m north and south, respectively, of plot center ( Figure 5 2). Height and basal diameter were measured for each shrub and sapling. Groundcover was sampled within four 1 1 m quadrats loc ated at the four cardinal directions and 4 m from plot center ( Figure 5 2). The north and south groundcover quadrats were nested within the shrub belt transects. For each quadrat, percent cover of herbaceous plants, grasses, vines, as well as shrubs and t rees less than 0.5 m in height, were quantified using ocular estimation from a vertical perspective. Shrub and tree seedling cover (<0.5 m) was further classified by species. Percent cover of each group was estimated regardless of overlap across groups, t herefore it was possible for cover to exceed 100%. Percent cover of litter or bare ground was estimated where vegetation cover did not occur. To dete rmine the effects of mowing on vegetation dynamics, vegetation measurements were compared across time sinc e treatment (TST), including pre treatment, within each of the stand types (mature, mature burned, plantation) in the buffer treatment. Tree (>0.5 m) density, basal area, and diameter (QMD) were
124 compared across TST (pre, post, 2 yr post) and stand type us ing a repeated measures analysis of variance (ANOVA), with TST as the within subjects variable and plot as the subject. Analyses included both fixed effects and their interaction. Using all tree s pooled, within stand types, diameter and height distributi ons were also created and compared between pre and post treatment. To evaluate treatment effects on the understory strata, density of saw palmetto, shrubs (including saw palmetto), and small trees (<2.5 cm DBH), as well as species richness of shrubs and small trees, pooled within plots, were compared across TST and stand type using repeated measures ANOVA. Mean density of shrubs and small trees, by species, were determined for each TST within stand types. Species that rarely occurred were not included. Percent groundcover by cover type (shrubs <0.5 m in height, tree seedlings <0.5 m in height, herbs, grasses, vines, litter, and bare ground) and species richness of groundcover shrubs and trees were each compared across TST and stand type using repeated measures A NOVA. For all analyses, statistical significance was test ed at the =0.05 level, and the Tukey Kramer post hoc comparison of the means test was used to determine diffe rences amongst groups. Each ANOVA was conducted as a within subjects analysi s with TST as the within subject variable and each plot as the subject. When model assumptions were not met, data were log or square root transformed to meet assumptions. In circumstances where occurrences were too rare for GLMANOVA (understory small tre e density and herb groundcover), the Chi Square test was used to determine if occurrences across levels of fixed effects were unlikely to have occurred at random.
125 To evaluate t he effects of both mowing and burning on vegetation dynamics, vegetation sampling was conducted following treatments in the experimental blocks. Mow and m ow+burn treatments were mow ed in August 2010 and all treatments, including controls, were sampled in October 2010. Mow+burn and burn only treatments were burned the last wee k of February 2011 ( Ch 4 ) and subsequently sampled in March/April, 2011. Vegetation sampling was then conducted again at one year following burning in March 2012. Three sampling plots were systematically located within each treatment 50 m from the edge o f each treatment, to avoid possible edge effects, and 50 m apart ( Figure 5.X). The same vegetation sampling technique was used as conducted in the buffer, except that saw palmetto cover (%) was also quantified over the entire 8 m radius plot (201 m 2 ), usi ng ocular estimation, at all sampling periods. Tree density, basal area, and QMD were compared across treatments (mow, mow+burn, burn only, control), within each sampling period. Saw palmetto, shrub (including saw palmetto), and small tree (<2.5 cm DBH) density were compared across treatments, within sampling period, as well as saw palmetto cover and species richness of understory shrubs and small trees pooled. Groundcover (%), by cover type (shrubs <0.5 m, tree seedlings <0.5 m, herbs, grasses, vi nes, litter, and bare ground), were compared across treatment s within sampling period, as well as species richness of ground cover shrubs and trees pooled. All comparisons were conducted using a GLMANOVA, except where occurrences were rare (understory sm all trees, herb and vine groundcover) where the Chi Square test was used, as above. Microclimate and Fuel Moisture Dynamics M icroclimate a nd moisture dynamics in surface litter and shrubs were evaluated within the experimental block treatments. At each plot location, between October 2010
126 (2 months post mowing) and January 2012, air temperature and relative humidity were record ed every 30 minutes using EasyLog EL USB 2 data logger s (DATAQ Instruments, Inc. Akron, OH) located at plot center 1 m above the g round, and soil temperature was r ecorded every 30 minutes using Watermark Soil Moisture Sensors, located 10 cm beneath the mineral soil, attached to Watchdog 450 data logger s (Spectrum Technologies, Inc. Plainfield, IL) Surface litter moisture content an d foliar moisture content of shrubs were sampled at each plot location every 3 4 weeks between June 2011 (14 weeks after burning) and March 2012. Fuel moisture content (FMC) sampling was initiated at 2.5 months following burning treatments to allow litter input and reestablishment of shrubs in the burned sites. Burning in the burn only and mow+burn treatments resulted in 100% area burned with high consumption of surface litter and understory shrubs ( Ch 4 ). Surface litter was collected at two locations wi thin each plot and pooled and foliar samples were clipped from two individuals of the dominant shrub species in each plot and pooled. Moisture samples were bagged, transported to the laboratory, oven dried at 65 C for 72 h, and gra vimetric FMC (water mass as a percentage of dry mass) calculated Air temperature, relative humidity, soil moisture, and soil temperature were each averaged by day and within two time period categories: day (08:00 19:59) and night (20:0 0 07:59). Each microclimate metric was then compared across treatments using a repeated measures general linear model analysis of variance (GLM ANOVA) in NCSS (Hintze 2008) with time since treatment (TST), by month, as the within subject factor, treatment (mow, mow+burn, burn, and control) as the between subject factor, and plot as the subject. Since burn treatments were conducted six months following mowing
127 treatments, planned comparisons were conducted to detect differences between mow (mow and mow+burn ) and un mowed (burn and control) treatments, pooled, prior to burning treatments and to detect if differences occurred between mow and mow+burn sites as well as burn and control sites prior to burning. Following burning treatments repeated measures analy sis was conducted with data separated into growing season (March August 2011) and dormant season (September January 2012) as seasonal effects were anticipated. Surface litter moisture and foliar moisture content of shrubs were compared across all treatmen ts immediately prior to ignition on the February burn day using an analysis of variance. Subsequently, FMC of litter and shrubs were each compared across all treatments between the June 2011 and May 2012 collections using a repeated measures GLM ANOVA as above, but with each sampling period as the within subject factor. Data were analyzed within season as above with growing season occurring between June and August 2011, and dormant season occurring between September and March 2012. Decomposition Decomposition of m owe d surface debris (fuels) was evaluated over a one year period. Recently mow ed surface fuels were collected from an adjacent site to that of the experimental block treatments, with similar overstory and understory vege tation. Surface debris was collected and transported to th e University of Florida Fire Science Laboratory. Fuels were oven dried at 50 C for one week. Fuels were then sorted into foliar litter (primarily saw palmetto) and woody debris separated into two size classes: <0.635 cm (1 h) and 0.635 2.54 cm (10 h). Woody fuels >2.54 cm were not collected in the field as there were only a few pieces found. Fuels, by type, were mixed by hand to reduce heterogeneity.
128 Decomposition bags (20 30 cm) were created from 2mm mesh screen material and sealed along edges with staples. 216 bags were filled with 50 g of foliar litter and 216 bags were fi lled with 50 g of 1 h woody fuels. In addition, 36 bags were filled with 10 h woody fuels, ranging from 23 to 58 g of fuel per bag. All bags were placed in laboratory conditions for 3 days, along with 10 birch ( Betula papyrifera Marsh.) medical grade ton gue depressors ("blanks") to estimate lab fuel moisture content (FMC). All decomposition bags, and blanks, were weighed prior to transporting bags to the field. Blanks were oven dried at 65 C for 72 h and weighed to back calculate gravimetric FMC (Eq. 5 1). Within each mow ed (N=9) and control (N=9) plot, in the experimental treatment sites, litter (6 ea) and 1 h (6 ea) decomposition bags were placed on surface fuels in a grid pattern and anchored with pin flags. Litter and 1 h bags were placed in treatment sites in February 2011, approximately 6 months following m owing Since there were not enough 10 h woody fuels to create six decomposition bags per plot, one 10 h bag was placed in each plot, however they were not placed in plots until April 20 11 during the first collection of litter and 1 h bags. Litter and 1 h decomposition bags were collected at two month intervals, following initial placement, for 12 months (April 2011 through February 2012) for a total of six collection periods. 10 h ba gs were collected at the same time, but were only collected five times over a 10 month decomposition period because they were initially placed in the field two months following litter and 1h fuels. During collection, one litter bag and one 1 h bag was ran domly selected, from each plot, for destructive sampling. The one 10 h bag in each plot was also collected, but returned following weighing. All bags were
129 transported to the UF laboratory, air dried for 3 days, and weighed. While 10 h bags were weighed and returned to the field, litter and 1 h bags, with enclosed fuel, were weighed and then fuel was removed and bags reweighed to calculate fuel weight, exclusively. Bag weights, without fuel, were subtracted from initial bag weights, at the start of th e study, to determine initial fuel weight, exclusively. 10 h fuels were removed from bags at the end of the study (Feb 2012) and subtracted from initial fuel weights, as well as all collection period weights, to calculated fuel weights, exclusively. Fuel weights at each time period were divided into original weights to calculate the proportion remaining at each collection period throughout the study. For each fuel type, the proportion remaining, as a percentage, was compared across treatments (mow and co ntrol ) using a repeated measured ANOVA with collection period as the within subject factor, treatment as the between subject factor, and plot as the subject. While litter and 1 h decomposition bags consisted of the same mass and 10 h bags did not differ i n weight between treatments (P=0.197), differences in decomposition between mowed and control sites were sought to determine if shrub cover influenced decomposition since shrub recovery is rapid in this system and over longer periods decomposition rates ma y be influenced by this recovery. Soil Nutrients To ev aluate the effects of mowi n g and burning on soil nutrients, soils were sampled within the experimental block treatments prior to implementing burn treatments (Feb 2011) and then again at one year fol lowing treatment. Soil nutrient pools were sampled, across all treatments, on Feb 1, 2011 three weeks prior to burning. At each plot, two soil samples were extracted at 4 m from plot center, at each of the four cardinal directions, using a 2 cm diameter soil push probe
130 Samples were separated into 0 5 cm and 5 10 cm soil depths and the two samples at each cardinal direction, by depth, combined. Three of the four resulting subsamples, of each depth, were randomly selected and pooled for nutrient analysis and the fourth used to estimate soil bulk density. Samples for nutrient analysis were air dried in a laboratory, sieved to remove roots >2 mm, and homogenized. Samples were analyzed by Waters Agricultural Laboratories (Camilla, GA, US ) Soil pH and ca tion exchange capacity (CEC) were determined. Total P, a vailable P, exchangeable K, Mg, and Ca, as well as the base saturation of K, Mg, Ca, and H were determined from Mehlich 1 extractions analyzed on an ICAP spectrometer. T otal N was determined from Kj eldahl digestion, t otal C from acid digestion a nd percent organic matter from loss on ignition. Bulk density samples were oven dried at 105 C for 24 h and weighed. All soil nutrient data were compared across treatment and time since treatment using a within subjects GLMANOVA. Results Vegetation Dynamics in the Buffer Area Mowi n g in the buffer treatments reduced overstory tree density in all stand types (mature, mature burned, plantation), but only significantly reduced basal area in the mature stands (Table 5 1). While density did not statistically differ between pre and post m owing in the plantation stands, density was lower 2 year s following mowing Quadratic mean diameter (QMD) in mature and mature burned stands significantly increased, however QMD was not affected by mowing in plantation stands. When pooling all trees within stand types, reduction in tree density in both mature and mature/burned stands primarily occurred in smaller diameter trees (<20 cm DBH), while trees in the plantation stands were reduced across all diameter classes, however less
131 so at greater diameters ( Figure 5 3). Pre treatment DBH ranged between 2.5 to 59.7 cm, 2.6 to 51.8 cm, and 2.8 to 30.7 cm in the mature, mature/burned, and plantation stands, respectively (Table 5 2, Figure 5 3). Immediately following treatment, u nderstory shrub density (>0.5 m in height) was reduced by 90, 85, and 70 % in the ma ture, mature/burned, and plantation stands, re spectively, following mowi n g Saw palmetto density, exclusively, was reduced by 74, 66, and 77 % in mature, mature/burned, and plantations stands, respectively. By 16 months, s hr ub and saw palmetto densities had increased to levels that did no t differ from pre treatment values (Table 5.2, Figure 5 4). Density of understory small trees (<2.5 cm DBH) did not statistically differ across time since treatment (TST) or stand type according to Chi Square analysis. Small tree occurrence was rare, resulting in high variability (Table 5 2, Figure 5 4). Species richness of understory shrubs and small trees was reduced following mowing across all stand types, due to immediate loss of most species ( Figure 5 5), but did not differ from pre treatment values at 16 months following treatment (Table 5 2, Figure 5 4) with most species re emerging ( Figure 5 5) Pre treatment groundcover was dominated by litter in the mature and plantation stands (~80%), and by both litter an d shrubs (~50% ea) in the mature burned stands (Table 5 3, Figure ure 5 6). Shrub groundcover (<0.5 m in height) was not affected by mowing in mature and p lantation stands, however shrub cover w as l ess in the mature/burned stands at 16 months, but then did not differ from pre treatment values after 2 years. Grass cover, ranging from only 0 to 2.4%, was reduced following treatment in all stands but recovered to pre treatment values by 16 months. Herb cover was rare and the Chi Square analysis indicated a difference by stand type, but
132 not TST, however TST was marginal (P=0 077). Prior to treatment, h erb cover was only observed in mature stands, but 2 months following treatment, h erbs existed in all stands For up to two years, herb cover continued to be o bserved in the mature stands, but only once again in the mature/burned stands at 16 months and not at all in plantations Vines were greater in cover in mature (4.3%) and plantation (4.4%) stands compared to mature/burned stands (1.5%) prior to treatment Vines did not differ at 2 or 8 months following treatment in mature stands, but by 16 months vine cover had increased to 8.7%. Vines were not affected in mature/burned stands, but were greater in plantation stands at 24 months compared to 8 months. In mature stands litter cover was reduced from 81.0% pre treatment values to 68.3% at 16 months following treatment and even lower (55.0%) after 2 years. In mature/burned stands, litter cover increased to 69.4% following treatment, but then was reduced to 5 4.7% at 8 months and did not differ from pre treatment values after that. Litter cover was unaffected in plantation stands. Bare ground was rare across stand types and TST (<3%) and variation was high where it did occur, however Chi Square analysis revea led a TST affect (P=0.049). Species richness of shrub and tree seedling groundcover was highest in mature/burned stands compared to others (P=0.015) and richness only differed between the post treatment sampling (2 months) and the 16 and 24 months samplin g periods (P=0.007). Vegetation Dynamics in the Experimental Block Area In the experimental block treatments, overstory tree density, basal area, QMD, and tree height did not differ across treatments (mow, mow+burn, burn only, control) during any stage of sampling (post mow, post burn, 1 yr post burn) (Table 5 4). Average pre
133 treatment tree density, basal area, QMD, and height, across all treatments, were 35829.5 tph, 19.31.2 m 2 1 27.20.8 cm, and 22.10.5 m, respectively. Understory shrub density (>0.5 m in height) was lower in the mow treatments compared to all other treatments following mowing (P=0.049), while saw palmetto density and cover in the mow and mow+burn treatmen ts were both lower than the control and burn only treatments (P<0.001) (Table 5 5, Figure 5 7). Following burning treatments, shrub density was lower in both the burn only and mow+burn treatments (P<0.001), but the mow only treatment no longer differed fr om the control. Saw palmetto density was still lower in the mow only treatment compared to the control following burning, but lowest in both burn only and mow+burn treatments (P<0.001). Saw palmetto cover was lower in all treatments (mow, mow+burn, burn only) compared to the control, but mow+burn treatments were even lower than that of the burn only treatments (P<0.001). One year following burning, shrub density no longer differed amongst treatments (P=0.195), however both saw palmetto density ( P=0.034) and cover (P<0.001) were lower in both the mow only and mow+burn treatments compared to the control and burn only treatments, but did not differ between mow and mow+burn treatments or between burn only treatments and controls. Small understory trees (<2.5 cm DBH) were rare across treatments with no significant differences between treatments using a Chi Square analysis, during any sampling periods. Small trees were only observed in control and mow only treatments during all sampling periods (Table 5 5). Sh rub height ( 5 m) and saw palmetto height were lower in both mow ed treatments prior to burning when compared with controls Both shrub and saw palmetto
134 height were reduced following burning, but while shrub height was still lower in mow only sites compared to contr ols, saw palmetto height did not diffe r between mow only and controls. Saw palmetto height was lowest in mow+burn sites compared to all others after burning One year following burning, shrub heights were lower in mow and mow+burn sites com pared to burn o nly and controls. Burn only shrub heights were similar to those of controls, and there were not differences between mow only and mow+burn sites. One year following burning, s aw palmetto height, however, was lower in both mow and mow+burn sites compared t o controls, but did not differ from burn only sites. Understory species richness of shrubs and small trees only differed during post burn sampling where the burn only and mow+burn treatments were lower in species diversity than control and mow only treatm ents. On average, species richness was very low, totaling fewer than three species Groundcover was dominated by litter across all treatments during all sampling periods and did not differ in cover (%) across treatments during any sampling period (Table 5 6, Figure 5 8). Shrub groundcover (<0.5 m in height) was next in dominance and only differed by treatment during post burn sampling, where shrub cover was lowest in the mow+burn treatment, but did not differ from the burn only treatment. Average grass c over was approximately double in the mow and mow+burn sites compared to the control and burn only sites, during initial post mow ( pre burn ) sampling, but did not differ statistically due to high variation Following burning treatments, grass cover was low er in the burn only treatments than mow and mow+burn treatments, but did not differ from the control. One year following burning, average grass cover was 10.85.5% and 8.33.6% in the mow and mow+burn treatments, respectively,
135 compared to the 2.70.9% and 2.01.5% cover in the control and burn only treatments, respectively, however treatment effects were marginal (P=0.057). Herb and vine cover were both rare across treatments and a significant treatme nt effect could not be detected using the Chi Square an alysis during any sampling period. Prior to burning, bare ground was not observed in burn only treatments or controls, but was observed in both the mow and mow+burn treatments. Mow only treatments were higher in bare ground (3.91.8%) versus burn only an d control, but mow+burn treatments (0.80.6%) did not differ from any other treatment. Species richness of shrub and tree seedling groundcover did not differ across treatments prior to burning (4.3 5.7), but mow+burn treatments were lower in richness follo wing burning compared to the mow only treatments and controls. At 2 years following burning, shrub/tree richness again did not differ across treatments. Species richness of all groundcover plants, including shrubs (<0.5 m), trees (<0.5m), herbs, vines, a nd grasses, were only evaluated at one year following burning treatments and did not differ statistically, however marginal results (p=0.062) provided some evidence of higher richness in the mowed sites (mow 10.3 1.3, mow+burn 10.2 1.0) compared to the un mowed sites (control 6.7 0.7, burn only 8.2 0.8). Microclimate and Moisture Dynamics For up to six months following the August 20 1 0 mowing treatments, and prior to prescribed burning, relative humidity (1 m aboveground) was lower in the mowed sites versus the un mowed (P<0.001), but there were no differences between controls and the pre burn burn only sites (P=0.523) or between the mow only sites and pre burn mow+burn sites (P=0.659) ( Figure 5 9). Air temperatures (1 m aboveground) di d not differ among any treatments (P=0.979) prior to burning. Following the February 2011
136 burning treatments, relative humidity did not differ among treatments (P=0.515) during the March to August growing season, however air temperatures were lower in the mow+burn sites versus the controls (P=0.036) but with a significant interaction between treatment and time since treatment (P=0.004) A ir temperatures did not differ across treatments during the month of March. Between September and January 2011, 13 to 17 months following mowing treatments and 7 to 11 months following burning, no differences were detected in above ground air temperature (P=0.881) or relative humidity (P=0.477) among all fuels treatments. Soil temperatures, at 5 cm depth, did not differ between mowed and non mowed sites (P=0.645), between controls and pre burn burn only sites (P=0.799), or between mowed and pre burn mow +burn sites (P=0.831) during th e six months following m o wi n g and prior to burning ( Figure 5 10). Following the February 2011 burning treatments, growing season (Mar Aug) soil temperatures differed across all treatments (mow, mow+burn, burn only, control) (P=0.013), except during the month of March where mow only treatments and controls did not differ based on the interacti on between treatment and time (P<0.001). During the remainder of the growing season (April Aug) soil temperatures consistently ranked highest to lowest across mow+burn, burn only, mow only, and controls, respectively. Soil temperatures did not differ co nsistently acros s treatments from September and January 2011 (P=0.723), however a strong interaction between treatment and time (P<0.001) revealed that an influence of treatment on soil tempe ratures differed across time. I n September mow only and control sites did not differ in soil temperature, however burn only sites were higher than both non burned sites, and mow+burn sites were higher than all others. In October
137 burned sites (burn only and mow+burn) had higher soil temperatures than unburned sites (mo w only and co ntrols). I n November burn only and mow only sites were actually lower in soil temperature than controls, and in December burn only sites were lower than all other treatments. Differences of moisture content in both surface lit ter (P<0.001) and live shrub foliage (P<0.001) were detected across treatments during the 10 months of sampling (3 13 months following burning), however interactions between treatment and time (P<0.001) indicate that differences were not consistent through out the sampling period ( Figure ure 5 11). When separating the growing season (June August), litter moisture was low and the interaction between treatment and time was not significant (P=0.073). During this period, controls were the wettest (13.20.8%), m owed treatments drier (9.70.6%), and mow+burn (7.10.4%) and burned (7.00.3%), not differing, were the driest. During the remainder of the sampling period (September March), differences in litter moisture across treatments were not consistent (P<0.001), but were generally highest in controls, lower in the mowed treatments, and lowest in the burn only and mow+burn treatments. Burn only treatments do appear to have higher litter moisture than mow+burn only during the wettest months (Sep Dec). Also, diffe rences in litter moisture detected across all treatments were most pronounced during these wetter periods. Decomposition Decomposition, quantified as dry mass remaining as a percentage of original, did not differ between mow treatments and controls throug hout a year of decomposition of foliar litter (P=0.249) or 1h (<0.635 cm) woody fuels (P=0.386) ( Figure 5 12). Mass remaining of 10h (0.635 2.54 cm) woody fuels also did not differ between treatments
138 throughout 10 months of decomposition (P=0.438). Inter actions between treatment and time were not significant for all three fuel types, however there is some evidence that final litter mass was higher in the mow treatments after one year. Foliar litter mass remaining after one year of decomposition was 74.1 6.0%, 1h woody mass remaining after one year was 82.35.3%, and 10h woody mass remaining after 10 months of decomposition was 81.25.0%. Soil Nutrients 3 (0 5 cm) and 1.27 to 1.36 cm 3 (5 10 cm), did not differ across experimental treatments (mow, mow+burn, burn only, control) before burning (2011) or one year following burning (2012). Pre burn sampling was conducted si x months following mowing in the mow and mow+burn treatments. Soil pH ranged from 3.6 to 3.9 at 0 5 cm depths, and from 3.8 to 4.1 at 5 10 cm, but did not differ across treatments pre or 1 yr post burn. Cation exchange capacity (CEC), ranging from 7.20 to 8.65 (0 5 cm) and 3.79 to 5.47 (5 10 cm) meq 1 did not differ across treatments pre or post burn. Prior to burning, the only difference in soil nutrients was that exchangeable K within 0 5 cm was lower in mow+burn treatments 2 ) compared to controls 2) ) (P=0.037) All other nutrients were similar across treatments. The only soil nutrient difference between treatments one year after burning was that base saturation of H within 0 5 cm was lower in burn only treatment s (83.36%) compared to control s (90.39%) (P=0.047) Discussi on The immediate ecological effects of mechanical mastication ("mowing") in palmetto/gallberry pine flatwoods in this study were short lived, suggesting that this ecosystem may recover quickly from such treatments. Ecological attributes were
139 evaluated onl y up to 2 years following treatment, however, and there was also evidence that saw palmetto, where abundant, may be significantly reduced through mowing Findings indicate that while understory vegetation is significantly reduced following mowing recover y to p re treatment conditions occurs rapidly. Mastication is a fuels treatment methods aimed at altering only understory shrubs and small trees, while leaving the overstory intact (Kane et al. 2009). In this study, while tree density was reduced following mowing treatments, basal area was only reduced in mature stands that had gone unburned for several years, where abundant smaller diameter trees occurred. Even in mature stands burned within 5 years prior to mowing tree density was reduced >40%, but basa l area did not differ. Mature unburned stands had more small diameter trees, prior to mowing con tributing to greater basal area than in burned stands. Long term fire exclusion in southeastern US pine forests typically results in increases in tr ee densit y and basal area and restoration goals often include the removal of under and mid story hardwood trees Tree removal from mowing in this study resulted in a 73% reduction in hardwood density in the mature stands. Mechanical treatments are likely to be u sed to reduce hazardous fuel accumulation in the understory, but also for restoration purposes in mature pine flatwoods where fire has been excluded. Rapid shrub recovery following mowing in pine flatwoods is likely due to the sprouting ability of domina nt species in this ecosystem. Saw palmetto and gallberry are common understory shrubs in the lower coastal plain of the southeastern US (Hough and Albini 1978) and dominated the pine flatwoods in this study. This vegetation strata is the primary driver o f fire behavior and effects in flatwoods ecosystems and rapid
140 recovery of dominant species is common following burning (Hough and Albini 1978, Abrahamson 1984a&b, Brose and Wade 2002). Recovery of shrub and saw palmetto cover to pre treatment conditions m ay occur as rapid ly as 1 to 2 years following fire (Abrahamson 1984a, Brose and Wade 2002) and desired prescribed burning cycles, as a fuels treatment method, is typically < 5yrs. Mowing in the buffer treatments in this study indicated recovery of shrub a nd saw palmetto density as quickly as 16 months following treatment. In the experimental treatments, however, saw palmetto cover was much lower in mowed site s. While pre treatment data were not available in these treatments, reduction in saw palmetto cov er following mowing in a nearby 500 ha areal treatment was similar to differences observed between mowed and control sites in the experimental treatments (Ch 2). Both sites had similar soils as well as over and under stories prior to treatment. Saw palm etto cover may have been lower in mowed sites due to mechanical damage to meristematic tissue in horizontal stems Stems existed in the mow treatments that appeared damaged, and where continued growth of palmetto fronds did not occur following treatment. While other sprouting shrubs have underground storage organs, apical meristems of saw palmetto occur above ground. They may be damaged from mowing where mowing heads are operated close to the ground or especially in areas where stems are elevated as a re sult of poorly drained soils Burning in pine flatwoods resulted in recovery of shrub density to that of pre treatment densities regardless of mowing treatments. Shrubs were less dense in mow only sites prior to burning, but did not differ from other tr eatments, or controls, one year following burning. Shrub density was measured for shrubs >0.5 m in height and many
141 of the resprouting shrubs, e.g. gallberry, were likely not tall enough to be considered in the understory when pre burn sampling occurred in the mow only sites. While saw palmetto cover was reduced to near zero immediately following burning in burn and mow+burn treatments, cover was no different between mow and mow+burn or between burn only and controls one year later, however saw palmetto co ver remained lower in both mowed treatments compared to burn only and controls. Mowing on the ONF as a management tool is primarily being conducted as a pre treatment to alter fuel structure prior to reintroduction of frequent prescribed burning Mowing in pine flatwoods may have the potential for reducing saw palmetto cover, but following mechanical fuels treatments with burning may not result in additional deleterious effects to the understory. This study, however, only revealed treatment effects and v egetation recovery up to two years following mowing and one year following single prescribed burns. Additional study would reveal possible longer term effects, especially under varying mowing and burning regimes over time. Groundcover was largely unaffe cted by mowing or burning treatments in this study. Litter and shrubs dominated groundcover across all study sites and at all stages of treatment. In the buffer, mature stands that were recently burned had more shrub groundcover than unburned stands or y ounger pine plantations across all time since treatment sampling, but shrub cover was not affected by mowing across stand types. Litter cover only decreased after two years in the mature unburned sites where vine cover had increased. While vines were pre valent in these stands prior to mowing they were primarily above 1 m in height and not considered part of the groundcover. Vine cover was minimal in recently burned mature stands and did not increase in the
142 groundcover strata following mowing Pre treatm ent vine cover in the pine plantation was similar to mature unburned stands, but incre ases in post treatment cover were not statistically significant due to higher variation. Herbs and grasses were minor components of groundcover in this study and are like ly inhibited by dominance of understory shrubs (Lewis and Hart 1972, Abrahamson 1984). Grass cover was reduced immediately following treatment across all stand types in the buffer, but while average grass cover appeared to approximately double that of pr e treatment after 2 years in all stand types, differences were not significant due to high variation In addition, mowed sites in the experimental treatments had over twice the average grass cover than un mowed sites, but a high degree of variation resul ted in insignificant differences. One year post burning, however, grass cover of mow only sites was 4 times that of controls and mow+burn sites were 4 times that of burn only sites, and differences were marginally detected. Because grass cover was low in un mowed sites and variation of grass cover high in mowed sites, differences were likely undetectable, even after transforming data to meet model assumptions. Both study locations revealed some evidence that grass cover increased due to mowing and the re duction of saw palmetto cover in experimental mowed treatments may have resulted in open ground for grasses to establish or spread Bare ground was higher in mowed sites prior to burning in the experimental treatment blocks. While burning treatments were implemented in February in this study, flowering response in grasses occur more readily during growing season (May July) burns in f l atwood s communities (Abrahamson 1984). Glitzenstein et al. (2003) found that while more frequent burning in Ultisol flatwoo ds in South Carolina shifted communities from being woody to herbaceous dominated, they did not find
143 increase s in herbaceous cover in Spodosol flatwoods on northern Florida, but did observe reduced dominance in saw palmetto and slight increased importance of forbs and grasses. Future study may reveal that a reduction in saw palmetto cover following mowing treatments may increase grass cover over time, especially under frequent growing season burning regimes. 0. 5 m) an d understory small trees (<2.5 cm DBH, but 0. 5 m in height) was reduced following mowing but recovered to pre treatment diversity after 16 months in the buffer treatments, yet did not differ between mowed and un mowed treatments in experimental blocks pr ior to burning. Burning did reduce understory shrub/tree species richness, but diversity recovered after one year. Understory diversity is quite low in this ecosystem, however, and richness was only evaluated at the 8 m 2 scale. While richness of underst ory shrubs and trees (both m) was initially reduced by burning, species richness of groundcover shrubs (<0.5 m) and trees (<0.5 m) was only reduced in mowed sites following burning, but sites were similar in diversity after one year. Abrahamson (1984b) observed an increase in diversity of woody plants following fire in flatwoods only after the first year and diversity was associated with an increase in overall evenness in abundance, not due to species ingress. Shrub and tree diversity recovers quickly in our flatwoods sites a nd is reflected in little change to groundcover following treatment and most notably by the fast recovery into the higher statured understory strata. When including all plant species into groundcover richness however, marginal evidence suggested here tha t mowing increased species richness one year following treatment, yet burning alone did not. While herb and grass cover were fairly low, compared to shrub cover in this study, there
144 was more variation in cover observed in mowed sites and the potential for increased diversity of all ground cover species following mowing may be a longer term effect of these treatments Reduction of saw palmetto cover may have attributed to such increases, where increased resources, especially light, could be allocated to mo re herbs and grasses. Increases in species diversity as a result of mowing in t his eco system will likely be associated with non woody species and longer term study may provide important insight into such an important effect. I ncreased herbs or grasses a t the cost of saw palmetto may have ecological benefi ts, but also potential negative effects In well drained Pinus palustris uplands, h igh understory plant diversity depend largely on frequent fire, wh ere herbs and grasses dominate groundcover ( Varner et al. 2000, Glitzentein et al. 2003 ). These fine fuels also a id in the ignitability in these uplands promoting frequent fire and further perpetuating their dominance. In mesic flatwoods, high fire frequency may also reduce shrub cover and enhanc e non woo dy herbaceous groundcover, however not to the extent of the more xeric uplands ( Glitzenstein et al. 2003 ) Restoration efforts in southern pine forests are often aimed at reducing shrubs and understory trees while increasing herbs and grasses (Varner et a l. 2000) M any wildlife species depend on this herbaceous groundcover as a food source and its loss has been associated with declines in several faunal species ( Van Lear et al. 2005 ) In flatwoods, w here shrubs dominate the understor y diversity is lower however shrubs are quite resilient to frequent fire (Abrahamson 1984 a ) and saw palmetto is an important food s ource for many wildlife species and individuals may be as old as 500 700 years (Tanner et al. 1996 ). While some loss of saw palmetto and slight increases in grasses and herbs from
145 understory mowing was observed in this study, further shifts in species composition may be an important management concern if such treatments are to repeated through a frequent management regime. Mowing and burning had minimal effects on understory microclimate, however the moisture regime in surface litter and live shrub foliage was significantly impacted. Although mow ed residues may have a mulching effect by slowing moisture loss ( Kreye et al. 2012 ), the drier lit ter in mowed sites compared to controls in this study was likely due to less saw palmetto cover and lower shrub heights increasing solar radiation and/or wind speed at the forest floor surface. It is unclear if a mulching effect in mowed residues attribut ed to higher litter moisture in mowed sites compared to burned sites because shrub heights and saw palmetto cover were higher in non burned sites, potentially resulting in similar effects on surface moisture. Although treatment effects on litter moisture were most pronounced during wetter months in the fall, effects on moisture during the drier sample periods in the summer were significant enough to have potential impacts on surface fire behavior. Although lower shrub cover and height may reduce fire int ensity during burning ( Ch 4 ), drier litter moisture during burning of mow ed sites may result in longer durations of lethal temperatures at the surface or in underlying soils (Ch 4). Potential tree mortality from burning in compact mow ed residues (Knapp et al. 2011) may result from long duration heating in these fuels (Kreye et al. 2011, Ch 4). When burning as a follow up treatment is not used, decomposition of mow ed residues may be an important factor regarding fuel dynamics in these treatments (cite). While higher moisture content, all else equal, tends to increase decomposition rates
146 (Enriquez et al. 1993, Prescott et al. 2004), decomposition of residues in this study did not differ between mowed sites and controls. Decomposition rates of surface litt er, however, were slightly higher than that observed in pine plantations in the region (Gholz et al. 1985). Dry litter mass remaining did not differ between mowed sites and controls during each 2 month collection period, except that there was some evidenc e of greater mass remaining in the mow treatment at the last collection (369 day s). Whether this difference would have been observe d under further decomposition is unknown since it was the end of the study. The average 74% remaining litter mass after one year was slightly less than the 85% litter mass remaining observed by Gholz et al (1985). Their study suggested that high lignin content and low P and N content in the needle dominated litter accounted for slower decomposition rates compared to other st udies, and not microclimate environments. In mow ed residues in our sites, saw palmetto leaves were likely a large proportion of surface litter due to pre treatment biomass, compared to the needle dominated litter in the Gholz et al. (1985) study. While l ignin content of litter in their study was 33 37%, saw palmetto lignin has been observed to be 18% (Pitman 1993). Also, C:N ratios in litter observed by Gholz et al. (1985) was 125 172, while C:N ratios of collected mow ed residues in a similar site near t his study, was 76.63.2 and 86.93.1 immediately post treatment and at one year following treatment, respectively (Kreye unpublished data). C:N ratios, like C:P ratios, are generally inversely related to decomposition rates (Enriquez et al. 1993) and the lower C:N ratios in litter following mowing may also attribute to faster decomposition rates compared to needle dominated litter in these pine forests. Understanding decomposition rates of mow ed residues will be important for predicting biomass, nutrient, carbon, and fuel
147 dynamics under varying treatment regimes in areas where burning may be difficult and mowing treatments used, exclusively, as a fuels treatment method. Soils nutrients were generally unaffected by mowing treatments or prescribed burning. Even where statistical differences between treatments occurred, there was no clear impact by either treatment exclusive ly. At 0 5 cm depths, K was only lower in mow+burn treatments compared to con t rols prior to burning a nd while base saturation of H at 0 5 cm was lower in burn only treatments compared with controls after burning, they didn't differ from mow or mow+burn sites. Lavoie et al. (2010) found increases in P, Ca, Mg, and K in the upper 5 cm of mineral soil 2 days following burning in a similar lo ngleaf pine slash pine flatwoods forest in northern Florida, but after one year P returned to pre burn levels and Ca, Mg, and K appeared to have had declined from that of elevated levels observed initially following burning. P and Mg levels were similar b etween sites in both studies prior to burning, however K was higher and Ca lower in our sites. Little research has been conducted on understory mechanical fuels treatments on soil properties, especially from mastication type of treatments Reduction of s oil moisture and soil respiration were observed following mastication in Sierra Nevada pine plantations, with mitigating effects on soil temperature changes (Kobziar and Stephens 2006). Moghaddas (2007) examined thinning treatments followed by burning n S ierra Nevada forests to increase N pools, exchangeable Ca, and pH. Consumption of duff (humus and fermentation layers) was high in their study, while little to no duff was consumed during burning in our study ( Ch 4 ). And following burning in the flatwood s site in the Lavoie (2010) study, there was a substantial reduction in C pools in these layers as well, indicating duff consumption during burning. Although litter
148 and understory shrub foliage was almost completely consumed, the remaining duff layer in ou r sites may have inhibited nutrient input, as a result of burning, into the mineral soil. Rhoades et al. (2012) found initial decreases in soil available N following addition of masticated mulch residues in Colorado coniferous forests, but after 3 5 years available N was greater in masticated sites versus controls. M asticated residues immediately add a nutrient pool to the forest floor however these nutrients are unavailable to plants until they are broken down and released in available forms And when residues are left on site, nutrients may be slowly released into the mineral soil over time One year following treatment may not be enough time to observe changes to mineral soils from the addition of these residues. The importance of nutrients in meta bolic processes means that plant foliage should have a greater proportion of plant nutrients contain within their tissues as compared to woody tissues, although C a i s an important component of wood cell walls (Chapin et al. 2002) The high proportion of f oliar litter in residues following mowing of palmetto/gallberry understories (Ch 2) may result in longer term inputs of nutrients into mineral soils when residues aren't burned. Especially where saw palmetto is abundant. Fire tends to alter N and P, two p rimary limited nutrients in this ecosystem, disproportionately due to the lower volatilization temperature of N versus P. Although available forms of soil N, such as NH 4+ may increase following burning i n palmetto flatwoods increases in available P (P O 4 3 ) may be greater thus reducing soil N:P ratio in the short term (Schafer 2012) These effects can attributed to similar changes in foliar N and P in saw palmetto following these same burns, where foliar N:P were also reduced (Schafer 2012). If mow ed residues are not burned, total N and P may both increase as litter is added to
149 the forest floor, but changes in nutrient availability to plants may rely on post treatment nutrient mineralization. Decomposition rates were a bit faster in these mow ed residu es than needle dominated litter in pine flatwoods plantations, but slower than many other areas (Gohlz et al. 1985). Mowing may alter soil nutrient s over time in areas where mowing is used a s a stand alone treatment option where frequently treating fuels will continue to add nutrients to the soil without the losses typically incurred during burning Longer term study is required to better understand potential long term effects of such treatments on soil properties in southeastern forests. While increa sed attention has been given to evaluating effects of fuels treatments in forest and shrub ecosystems, little has been conducted with regard to understory mastication, especially in the southeastern US Much of the research evaluating mastication treatmen ts has been conducted in the western US and studies often address implications on fuel properties and potential impacts of treatments in altering fire behavior (Bradley et al. 2006, Hoo d and Wu 2006, Kane et al. 2009 Kobziar et al. 2009, Battaglia et al. 2010, Kreye et al. 2011). Unforeseen consequences of burning in masticated treatments has been documented in shrublands (Bradley et al. 2006) and coniferous forests (Knapp et al. 2011) in California, where heavy loading of woody dominated residues result from mastication (Kane et al. 2009). Although compact, these woody dominated residues may result in long duration surface (Kreye et al. 2011) and soil (Busse et al. 2005) heating, and high fire severity (Bradley et al. 2006, Knapp et al. 2011) when burned Studies aimed at evaluating vegetation response to mastication is lacking, but is gaining attention in the west. Kane et al. (2010 ) found that mastication alone in a northern Sierra Nevada ponderosa pine forest reduced midstory
150 vegetation, but did not affect understory diversity. Follow up prescribed burning, however, did increase diversity of both native and non native species. Potts and Stephens (2009), in contrast, found greater abundance of non native invasive species in masticated sites versus b urned sites in chamise ( Adenostoma fasciculatum ) dominated chaparral in northern California, but had had no affect on overall diversity. Increased cover of forbs and grasses in masticated pinyon juniper woodlands has been observed, but no differences in s hrubs (Ross et al. 2012). While plant responses to understory mastication treatments will likely vary across ecosystems, southeastern pine forests dominated by saw palmetto/gallberry understories is a unique ecosystem in its rapid recovery of vegetation c omposition and structure following mowing and burning. The potential reduction in saw palmetto was the one major effect observed in this study. The resiliency of palmetto/gallberry pine flatwoods to mastication ("mowing") and burning treatments was stri king in this study. Effects to vegetation, microclimate, moisture regimes, and edaphic features were either minimal or short lived. Other than reduction in saw palmetto, effects of a single mechanical treatment alone or followed by prescribed burning is minimal, at least in the short term. If mastication type of fuels treatments are to be used as an alternative to prescribed burning, treatment regimes are likely to occur on a frequent basis due to rapid recovery of shrubs. Evaluating ecological response to repeated treatments over time will be imperative if such treatment regimes are to be implemented. This study suggests that using mastication as pre treatment to follow up prescribed burning is a feasible option with minimal effects to the ecological i ntegrity of this ecosystem.
151 Table 5 1 Tree density, basal area, and quadratic mean diameter (QMD) following mowing treatments in pine flatwoods of northern Florida, USA. Stand Type mature mature burned plantation Stand Type TST a Stand Type TST -------------------p value ---------------------Tree Density -------------trees h a 1 ------------<0.001 <0.001 <0.001 Pre Treatment 941 (179) A 365 (36) A 1 0 20 (185) a Post Treatment 327 (58) B 216 (30) B 804 (82) ab 2yrs Post Treatment 290 (46) B 216 (30) B 713 (71) b Basal Area -----------m 2 ha 1 -----------0.029 <0.001 0.043 Pre Treatment 28.3 (3.3) A 17.9 (2.2) A 34.0 (5.9) A Post Treatment 23.2 (2.9) B 17.3 (2.4) A 27.5 (2.2) A 2yrs Post Treatment 23.4 (2.8) B 18.2 (2.3) A 26.3 (2.5) A QMD -------------cm -------------0.004 <0.001 < 0.001 Pre Treatment 21.8 (1.7) A 25.6 (2.1) A 20.7 (0.4) A Post Treatment 32.2 (2.0) B 32.8 (2.2) B 21.0 (0.5) A 2yrs Post Treatment 33.6 (1.7) B 33.9 (2.2) B 21.8 (0.6) A a Time since t reatment Note: Values sharing letters within rows are not statisti cally different (Tukey
152 Table 5 2 Density and species richness of understory shrubs and small trees following mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. Stand Type Mature Mature/Burned Plantation Stand Type TST a Stand Type TST --------------------individuals ha 1 ----------------------------------------p value -----------------Saw Palmetto b 0.060 <0.001 0.855 Pre Treatment 4038 (1032) A 6094 (1303) A 2708 (936) A 2 months 1042 (431) B 2083 (551) B 625 (280) B 8 months 909 (380) B 2778 (972) B 833 (417) B 16 months 2604 (877) A 4844 (1386) A 2292 (879) A 24 months 2604 (820) A 4375 (1398) A 2083 (768) A Shrubs c Pre Treatment 42308 (7093) A 69821 (16555) A 22500 (5995) A 0.049 <0.0 01 0.187 2 months 4167 (861) B 10694 (3126) B 6667 (2534) B 8 months 5227 (2177) B 15278 (2809) B 6250 (3354) B 16 months 37500 (9606) A 50417 (7321) A 22083 (5017) A 24 months 34545 (8013) A 43393 (6224) A 24792 (4113) A Small Trees d 0.455 0.427 na Pre Treatment 577 (304) 2083 (1021) 417 (264) 2 months 208 (208) 0 (0) 833 (833) 8 months 341 (244) 0 (0) 208 (208) 16 months 625 (288) 556 (303) 208 (208) 24 months 833 (444) 1528 (1098) 625 (280) Species Richness e --------------------species 8m 2 -------------------0.001 <0.001 0.543 Pre Treatment 3.5 (0.3) A 4.9 (0.5) A 2.5 (0.3) A 2 months 1.3 (0.1) B 2.6 (0.4) B 1.7 (0.2) B 8 months 1.5 (0.3) B 2.1 (0.3) B 1.5 (0.3) B 16 months 2.9 (0.6) A 3.8 (0.2) A 2.7 (0.4) A 24 months 3.1 (1.4) A 4.1 (1.4) A 2.8 (0.2) A Note: Values sharing letters within columns are not statistically different (Tukey a Time since treatment b Saw Palmetto ( Serenoa repen s ) individuals c All shrubs including saw palmetto d Trees <2.5 cm DBH e All understory shrubs and trees (<2.5 cm) pooled Chi Square test used to test main effects only due to rarity of occurrence of small trees across all stand types and time TST (time since treatment).
153 Table 5 3 Percent groundcover, by vegetation type, and species richness of shrubs (<0.5 m) and tree saplings (<0.5 m) following mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA. Stand Type Ma ture Mature/ Burned Plantation Stand Type TST a Stand Type TST -------------% --------------p value Shrub Cover b <0.001 <0.001 0.004 Pre Treatment 18 (3) A 57 (8) A 17 (3) A 2 months 18 (3) A 42 (4) AB 17 (3) A 8 months 20 (3) A 43 (2) AB 18 ( 4) A 16 months 18 (3) A 34 (3) B 12 (2) A 24 months 25 (4) A 53 (6) A 17 (3) A Grass Cover 0.265 <0.001 0.887 Pre Treatment 0.9 (0.4) AB 1.0 (0.4) AB 0.6 (0.4) AB 2 months 0.0 (0.0) C 0.9 (0.5) C 0.0 (0.0) C 8 months 0.3 (0.2) BC 1.7 (1.0) BC 0.2 (0.2) BC 16 months 1.4 (0.7) A 2.4 (0.6) A 1.9 (1.4) A 24 months 1.7 (0.6) A 1.9 (0.8) A 1.5 (0.9) A Herb Cover <0.001 0.077 na Pre Treatment 1.7 (0.8) 0.0 (0.0) 0.0 (0.0) 2 months 0.4 (0.2) 0.1 (0.1) 3.8 (3.7) 8 months 0.2 (0.2) 0.0 (0.0) 0.0 (0.0) 16 months 1.1 (0.6) 0.1 (0.1) 0.0 (0.0) 24 months 0.1 (0.1) 0.0 (0.0) 0.0 (0.0) Vine Cover c 0.014 <0.001 0.001 Pre Treatment 4.3 (2.1) A 1.2 (0.5) A 4.4 (1.5) AB 2 months 3.8 (1.2) A 2.0 (1.3) A 4.4 (2.6) AB 8 months 2.2 (0.6) A 1.3 (0.8) A 3.5 (2.0) A 16 months 8.7 (2.6) B 1.0 (0.4) A 10.8 (4.2) AB 24 months 16.4 (5.1) B 1.9 (1.2) A 13.0 (5.2) B Litter <0.001 <0.001 0.001 Pre Treatment 81.0 (3.1) A 53.6 (5.5) AC 82.6 (2.5) A 2 months 79.4 (2.5) AB 69.4 (3.9) B 79.5 (5.7) A 8 months 76.3 (3.2) AB 54.7 (3.4) AC 78.0 (4.4) A 16 months 68.3 (4.2) B 61.3 (3.7) ABC 75.0 (2.7) A 24 months 55.0 (5.8) C 45.8 (4.8) ACD 70.7 (2.9) A Bare Ground 0.714 0.049 na Pre Treatment 0.1 (0.1) 2.5 (2.2) 0.3 (0.2) 2 months 2.4 (1.0) 3.0 (1.5) 0.6 (0.5) 8 months 0.3 (0.3) 0.0 (0.0) 0.2 (0.2) 16 months 1.7 (1.3) 0.6 (0.4) 0.1 (0.1) 24 mont hs 0.5 (0.4) 0.4 (0.3) 0.7 (0.3) Species Richness d ----------------2 ---------------0.015 0.007 0.511 Pre Treatment 4.2 (0.5) AB 5.1 (0.4) AB 3.7 (0.3) AB 2 months 3.3 (0.5) A 5.8 (0.2) A 3.0 (0.6) A 8 months 3.6 (0.6) AB 5.1 (0.5) AB 3.7 (0.5) AB 16 months 4.4 (0.7) B 6.1 (0.4) B 4.2 (0.3) B 24 months 4.5 (0.8) B 6.2 (0.6) B 4.0 (0.6) B Note: Values sharing letters within columns are not statistically different (Tukey Kramer Table 5 3. Continued
154 a Time since treatmen t b Shrubs <0.5 m in height c Vines <1 m above the ground d Groundcover of shrubs and tree seedlings (both <0.5 m in height) Chi Square test used to test main effects only due to rarity of occurrence of small trees across all stand types and time TST (ti me since treatment).
155 Table 5 4 Tree density, basal area, and quadratic mean diameter (QMD) across experimental treatments following mowing and burning in pine flatwoods of northern Florida, USA. Treatment Status Post Mow Post Burn a 1 yr Post Burn Tree Density -------------t rees h a 1 ------------Control 426 (64) A 419 (59) A 419 (59) A Burn Only 359 (60) A 359 (58) A 326 (45) A Mow 326 (69) A 332 (69) A 332 (69) A Mow+Burn 337 (47) A 337 (47) A 332 (45) A Basal Area ----------m 2 ha 1 -----------Control 19.0 (2.3) A 19.1 (2.3) A 18.9 (2.2) A Burn Only 16.6 (1.4) A 16.7 (1.5) A 15.9 (1.4) A Mow 19.2 (2.4) A 19.5 (2.4) A 19.3 (2.5) A Mow+Burn 22.3 (3.4) A 22.5 (3.4) A 21.5 (3.3) A QMD -------------cm -------------Control 24.3 (1.0) A 24.5 (1.0) A 24.4 (1.0) A Burn Only 26.0 (2.2) A 25.7 (1.9) A 25.9 (1.9) A Mow 29.0 (1.3) A 29.0 (1.3) A 28.7 (1.2) A Mow+Burn 28.9 (1.4) A 29.0 (1.4) A 28.5 (1.3) A Tree Height --------------m --------------Control 22.2 (0.3) A 21.7 (0 .4) A 22.5 (0.5) A Burn Only 25.2 (1.6) A 20.4 (1.3) A 23.0 (1.5) A Mow 22.7 (0.5) A 22.5 (0.8) A 22.2 (0.8) A Mow+Burn 22.3 (1.0) A 22.4 (1.0) A 24.1 (1.3) A Note: Values sharing letters within columns are not statistically different (Tukey Kramer Test, a 6 months post mow.
156 Table 5 5 Density and species richness of understory shrubs and small trees, and percent cover of saw palmetto, across experimental mowing and burning treatments in pine flatwoods of northern Florida, USA. Treatment Statu s Post Mow Post Burn f 1 yr Post Burn --------------------individuals ha 1 ---------------------Saw Palmetto a Control 12500 (2148) A 12857 (2451) A 11964 (1786) A Burn Only 12222 (1246) A 139 (139) B 12639 (1537) A Mow 3333 (1755) B 4861 (2209 ) C 5694 (1681) B Mow+Burn 3056 (1781) B 0 (0) B 5694 (2448 ) B Shrubs b Control 28393 (5521) A 28750 (8192) A 26071 (5916) A Burn Only 30417 (4709) A 139 (139) B 39444 (9254) A Mow 16528 (2355) B 26528 (5519) A 31111 (4950) A Mow+Burn 23611 (3833) A 0 (0) B 48333 (8125) A Small Trees c Control 357 (231) A 357 (231) A 357 (231) A Burn Only 0 (0) A 0 (0) A 0 (0) A Mow 694 (694) A 972 (828) A 972 (828) A Mow+Burn 0 (0) A 0 (0) A 0 (0) A Shrub Height -----------------------------m ----------------------------Co ntrol 1.19 (0.05) A 1.17 (0.05) A 1.09 (0.04) A Burn Only 1.16 (0.04) A 0.06 (0.06) B 1.00 (0.05) A Mow 0.66 (0.03) B 0.69 (0.03) C 0.83 (0.05) B Mow+Burn 0.65 (0.02) B 0.00 (0.00) B 0.67 (0.02) B Saw Palmetto Cover d ------------------------------% ----------------------------Control 47.4 (8.1) A 62.1 (8.9) A 58.6 (9.3) A Burn Only 51.7 (8.7) A 19.4 (3.7) B 51.7 (7.0) A Mow 10.6 (2.7) B 11.1 (1.6) BC 21.1 (3.8) B Mow+Burn 8.9 (2.5) B 3.6 (1.0) C 13.3 (4.0) B Saw Palmetto Height d -----------------------------m ----------------------------Control 1.06 (0.06) A 1.04 (0.08) A 1.16 (0.04) A Burn Only 1.09 (0.05) A 0.51 (0.05) B 0.93 (0.0 7) AB Mow 0.68 (0.10) B 0.84 (0.04) A 0.89 (0.03) B Mow+Burn 0.69 (0.10) B 0.30 (0.05) C 0.71 (0.09) B Species Richness e --------------------species 8m 2 -------------------Control 2.9 (0.3) A 2.4 (0.2) A 2.0 (0.2) A Burn Only 2.9 (0.4) A 0.1 (0.1) B 2.1 (0.2) A Mow 2.2 (0.2) A 2.8 (0.3) A 2.8 (0.5) A Mow+Burn 2.3 (0.3) A 0 .0 (0 .0 ) B 2.2 (0.2) A Note: Values sharing letters within columns are not statistica lly different (Tukey Kramer a Saw Palmetto ( Serenoa repens ) individuals. b c Trees <2.5 cm DBH.
157 Table 5 5. Continued d Over entire 8 m radius plot (201 m 2 ), hei ght includes all palmetto regardless of height. e All understory shrubs and trees (<2.5 cm) pooled. f 6 months post mow. Chi Square test used to test effects due to rarity of occurrence.
158 Table 5 6 Percent groundcover, by vegetation type, and species richness of shrubs (<0.5 m) and tree saplings (<0.5 m) across experimental mowing (mowing) and burning treatments in pine flatwoods of northern Florida, USA. Treatment Status Post Mow/ Pre Burn Post Burn f 1 yr Post Burn -------------------% -------------------Shrub Cover a Control 27.7 (6.0) A 26.9 (3.9) A 23.1 (7.7) A Burn Only 19.6 (3.7) A 13.6 (1.9) AB 24.4 (2.5) A Mow 16.2 (2.0) A 26.0 (3.6) A 19.1 (2.2) A Mow+Burn 21.4 (2.7) A 8.0 (2.7) B 23. 2 (3.5) A Grass Cover Control 1.8 (1.0) A 4.1 (1.1) AB 2.7 (0.9) A Burn Only 1.5 (0.9) A 0.9 (0.7) B 2.0 (1.5) A Mow 4.6 (2.2) A 5.4 (2.4) A 10.8 (5.5) A Mow+Burn 6.4 (3.4) A 6.6 (2.9) A 8.3 (3.6) A Herb Cover Control 0.0 (0.0) A 0.0 (0.0) A 0.0 (0.0) A Burn Only 0.2 (0.1) A 0.0 (0.0) A 0.0 (0.0) A Mow 0.8 (0.8) A 0.2 (0.1) A 0.2 (0.1) A Mow+Burn 0.0 (0.0) A 0.1 (0.1) A 0.2 (0.2) A Vine Cover b Control 2.0 (2.0) A 0.7 (0.7) A 0.5 (0.5) A Burn Only 0.3 (0.3) A 0.1 (0.1) A 1.7 (1.5) A Mow 6.3 (3.6) A 4.0 (2.3) A 2.4 (1.3) A Mow+Burn 3.6 (1.6) A 0.2 (0.1) A 0.3 (0.2) A Litter Control 68.7 (5.8) A 67.9 (3.8) A 74.2 (7.3) A Burn Only 78.0 (3.9) A 42.9 (6.1) A 62.7 (3.9) A Mow 67.5 (4.1) A 62.4 (4.6) A 66.1 (5.5) A Mow+Burn 68.1 (4 .1) A 51.1 (13.8) A 66.2 (4.0) A Bare Ground Control 0.0 (0.0) A 0.0 (0.0) A 0.1 (0.1) A Burn Only 0.0 (0.0) A 41.9 (5.2) B 10.1 (2.5) B Mow 3.9 (1.8) B 3.1 (1.2) A 1.3 (0.4) C Mow+Burn 0.8 (0.6) AB 33.9 (11.2) B 5.7 (0.8) B Shrub/Tree Richness c Control 4.7 (0.7) A 5.6 (0.5) A 5.7 (0.5) A Burn Only 4.3 (0.4) A 4.2 (0.4) AB 5.7 (0.4) A Mow 4.6 (0.3) A 6.0 (0.4) A 6.0 (0.4) A Mow+Burn 5.7 (0.4) A 2.6 (0.7) B 5.8 (0.8) A Species Richness d Control 6.7 (0.7) A Burn Only 8.2 (0.8) A Mo w 10.3 (1.3) A Mow+Burn 10.2 (1.0) A Note: Values sharing letters within columns are not statistically different (Tukey Kramer Results of Chi Square test, marginal results of ANOVA (P=0.062) a Shrubs <0.5 m in height, b v ines < 1 m above the ground c shrubs (<0.5 m) and trees (<0.5) only d all groundcover plant species (shrubs, trees, herbs, grasses, vines) f 6 months post mow
159 Tabl e 5 7 Soil properties and nutrients across experimental mowing and burning treatments in pine flatwoods of northern Florida, USA. Bulk Density pH g cm 3 C B M M+B C B M M+B 0 5cm 2011 1.07 (0.05) A 1.02 (0.06) A 1.07 (0.06) A 0.94 (0.02) A 3.6 (0.04) A 3.6 (0.05) A 3.8 (0.03) A 3.8 (0.11) A 2012 1.10 (0.04) A 0.96 (0.03) A 1.06 (0.05) A 1.04 (0.06) A 3.8 (0.06) A 3.9 (0.03) A 3.9 (0.04) A 3.9 (0.16) A 5 10cm 2011 1.27 (0.07) A 1.36 (0.04) A 1.35 (0.04) A 1.30 (0.03) A 3.9 (0.06) A 3.8 (0.04) A 3.9 (0.03) A 3.9 (0.06) A 2012 1.29 (0.03) A 1.33 (0.03) A 1.35 (0.03) A 1.36 (0.04) A 4.1 (0.07) A 4.0 (0.03) A 4.1 (0.05) A 4.0 (0.05) A CEC Exchangeable K meq 100g 1 g 2 C B M M+B C B M M+B 0 5cm 2011 7.88 (0.41) A 8.18 (0.43) A 7.20 (0.43) A 7.54 (0.53) A 1.04 (0.13) A 0.99 (0.07) A B 0.87 (0.05) A B 0.72 (0.06) B 2012 8.17 (0.27) A 8.65 (0.30) A 8.34 (0.36) A 8.07 (0.41) A 0.95 (0.05 ) A 0.91 (0.05) A 0.99 (0.09) A 0.91 (0.06) A 5 10cm 2011 3.79 (0.37) A 5.37 (0.58) A 4.65 (0.41) A 4.59 (0.65) A 0.52 (0.07) A 0.67 (0.05) A 0.79 (0.07) A 0.62 (0.08) A 2012 4.34 (0.33) A 5.47 (0.58) A 4.91 (0.37) A 4.73 (0.48) A 0.54 (0.09) A 0.72 (0.08) A 0.60 (0. 05) A 0.68 (0.06) A Exchangeable Mg Exchangeable Ca g 2 g 2 C B M M+B C B M M+B 0 5cm 2011 1.73 (0.17) A 2.61 (0.52) A 2.33 (0.33) A 1.54 (0.21) A 6.33 (1.12) A 6.71 (1.18) A 6.98 (0.84) A 5.50 (0.33) A 2012 1.69 (0.17) A 2.69 (0.38) A 2.22 (0.25) A 2.33 (0.36) A 5.16 (0.49) A 6.4 6 (0.51) A 5.77 (0.87) A 6.59 (0.88) A 5 10cm 2011 0.89 (0.11) A 1.73 (0.29) A 1.68 (0.21) A 1.27 (0.26) A 4.61 (0.79) A 5.71 (0.82) A 6.83 (0.74) A 5.83 (0.70) A 2012 0.83 (0.14) A 1.45 (0.23) A 1.20 (0.24) A 1.26 (0.24) A 3.63 (0.44) A 4.51 (0.55) A 3.95 (0.64) A 5 .48 (0.77) A
160 Table 5 7 C ontinu ed Base Saturation (K) Base Saturation (Mg) % % C B M M+B C B M M+B 0 5cm 2011 0.63(0.06) A 0.61(0.03) A 0.62(0.03) A 0.53(0.04) A 3.43(0.25) A 4.84(0.56) A 4.91(0.31) A 3.69(0.45) A 2012 0.54(0.03) A 0.56(0.02) A 0.57(0.03) A 0.57(0.04) A 3.14(0.25) A 5.28(0.65) A 4.21(0.41) A 4.66(0.70) A 5 10cm 2011 0.59(0.09) A 0 .50(0.04) A 0.65(0.05) A 0.59(0.08) A 3.10(0.24) A 3.84(0.34) A 4.42(0.30) A 3.49(0.50) A 2012 0.51(0.09) A 0.51(0.04) A 0.51(0.05) A 0.60(0.11) A 2.49(0.39) A 3.38(0.52) A 2.92(0.43) A 2.82(0.56) A Base Saturation (Ca) Base Saturation (H) % % C B M M+B C 3 B M M+B 0 5cm 2011 7.35(0.97) A 7.76(0.85) A 9.20(0.86) A 8.39(0.76) A 88.59(1.14) A 86.80(1.27) A 85.28(0.91) A 87.26(1.09) A 2012 5.91(0.69) A 8.82(0.91) A 6.48(0.77) A 7.86(0.92) A 90.39(0.82) A 83.36(1.24) B 88.75(1.08) A B 83.79(1.54) A B 5 10cm 2011 9.42(0.92) A 7.89(0.70) A 10.83(0.66) A 10.25(1.46) A 86.86(1.16) A 87.77(0.84) A 84.11(0.93) A 85.73(1.74 ) A 2012 6.59(0.77) A 6.65(0.85) A 5.85(0.87) A 7.44(1.01) A 90.39(1.11) A 89.45(1.16) A 90.62(1.18) A 89.24(1.51) A Available P Organic Matter g 2 % C B M M+B C B M M+B 0 5cm 2011 0.38 (0.09) A 0.29 (0.03 A 0.28 (0.04) A 0.22 (0.03) A 2.77 (0.18) A 3.18 (0.29) A 2.50 (0.24) A 2.26 (0.26) A 2012 0.18 (0.03) A 0.28 (0.07) A 0.17 (0.02) A 0.21 (0.04) A 2.75 (0.20) A 3.34 ( 0.34) A 2.79 (0.22) A 2.64 (0.27) A 5 10cm 2011 0.27 (0.05) A 0.31 (0.03) A 0.45 (0.08) A 0.26 (0.04) A 0.88 (0.08) A 1.30 (0.16) A 1.07 (0.10) A 1.11 (0.21) A 2012 0.23 (0.05) A 0.31 (0.05) A 0.23 (0.04) A 0.23 (0.03) A 1.09 (0.12) A 1.26 (0.12) A 1.18 (0.12) A 1.1 1 (0.11) A
161 Table 5 7 C ontinued Total P Total N ppm % C B M M+B C B M M+B 0 5cm 201 1 52.60(2.34) A 57.04(4.17) A 50.31(2.70) A 52.38(3.59) A 0.20(0.00) A 0.21(0.01) A 0.20(0.01) A 0.21(0.02) A 2012 34.49(3.20) A 42.86(3.83) A 38.28(2.99) A 34.44(3.93) A 0.21(0.01) A 0.25(0.02) A 0.23(0.01) A 0.22(0.01) A 5 10cm 2011 32.02(1.62) A 37.41(1.78) A 36.54(2.09) A 33.85(1.78) A 0.15(0.01) A 0.16(0.00) A 0.16(0.01) A 0.16(0.01) A 2012 14.76(1.77) A 18.22(1.67) A 16.70(1.36) A 16.36(1.16) A 0.17(0.01) A 0.18(0.01) A 0.18(0.01) A 0.18(0.01) A Total C % C B M M+B 0 5cm 2011 2.08 (0.25) A 2.46 (0.38) A 1.77 (0.11) A 1.88 (0.36) A 2012 2.24 (0.22) A 3.10 (0.41) A 2.60 (0.25) A 2.18 (0.28) A 5 10cm 2011 0.50 (0.06) A 0.83 (0.14) A 0.6 7 (0.08) A 0.71 (0.15) A 2012 0.60 (0.08) A 0.89 (0.11) A 0.71 (0.11) A 0.88 (0.16) A Note: Value s sharing letters within rows by soil property are not statistically different (Tukey between treatments: control (C), burn only (B), mow only (M) mow followed by burning (M+B) Samples taken 6 months following mowing, 1 week prior to burning (2011); and 1 year following burning (2012)
162 Figure 5 1 Fuels treatments used for the study of the ecological effects of understory mowing in pine flatwoods of the Osceola National Forest (ONF) in northern Florida, USA. Three treatment areas are shown. 1) a 100 m wide and 6 km (60 ha) buffer masticated ("mowed") in 3 stand types: mature pine (ca. 80 yrs old), mature pine recently burned (5 yrs prior to mowing ), and young pine plantation (28 yrs old); 2) a 500 ha areal treatment (sampling plots exist in mature pine only); and 3) three e xperimental blocks each with the following treatments: mow, mow followed by burning, burn only, and control.
163 Figure 5 2 Vegetation sampling plots systematically allocated within a fuels treatment buffer on the Osceola National Forest in northern Florid a, USA. Plots were located at the center of delineated stand types (mature, mature burned, young plantation).
164 Figure 5 3 Tree height and diameter distributions pre and post treatment following mowing in 3 stand types (mature, mature/bu rned, plantation) in pine flatwoods in northern Florida, USA.
165 Figure 5 4 Density and species richness of understory shrubs and small trees following mechanical mowing of understory shrubs and small trees in pine flatwoods of northern Florida, USA.
166 Figure 5 5 Density by species of understory shrubs (left) and trees <2.5 cm DBH (right).
167 Figure 5 6 Groundcover (%), by cover type, and species richness of shrubs (<0.5m in height) and tree seedlings (<0.5 m in height) fo llowing mowing in 3 stand types (mature, mature/burned, plantation) in pine flatwoods in northern Florida, USA.
168 Figure 5 7. Density and species richness of understory shrubs and small trees across experimental mowing and burning treatment s in pine flatwoods of northern Florida, USA.
169 Figure 5 7 Continued
170 Figure 5 8 Percent groundcover, by cover type, and species richness of shrubs (<0.5 m) and tree saplings (<0.5 m) across experimental mowing and burning treatments in pine flatwoods in northern Florida, USA.
171 Figure 5 9 Average temperature (above) and relative humidity (below) across 3 fuels treatments (burn, mow, mow+burn) and controls up to 17 months following mowing treatments conducted in August 2010. Burning tre atments were conducted in Feb 2011, six months following mowing.
172 Figure 5 10 Average soil temperature, at 5 cm depth, across 3 fuels treatments (burn, mow, mow+burn) and controls up to 16 months following mowing treatments conducted in August 2010. Burning treatments were conducted in Feb 2011, six months following mowing.
173 Figure 5 11 Moisture content (%) of surface litter (left) and live shrub foliage (right) across fuels treatments (mow, mow+burn, burn only), and controls, in mature pine flat woods of northern Florida, USA. Moisture content sampled every 3 to 4 weeks between June 2011 and March 2012. Inserts indicate moisture content differences by treatment during the driest season.
174 Figure 5 12 Comparison of decomposition of surface litt er and surface woody debris (1h: <0.625 cm; 10h: 0.625 2.54 cm) created from mowing of saw palmetto and gallberry dominated understory of mature pine flatwoods of northern Florida, USA between mowed treatments and un mowed controls. All material collected for decomposition study were derived from understory mowing, however decomposition rates evaluated in un mowed controls was to determine if shrub cover influenced decomposition since shrub recovery following mowing is rapid. No differences in decompositi on were detected between treatments across any of the fuel types: litter (P=0.249), 1h woody (P=0.386), 10h woody(P=0.438).
175 CHAPTER 6 CONCLUSIONS The research presented here provides much needed insight into the effectiveness and effects of a n increasingly utilized mechanical fuels treatment method in a common forest ecosystem of the southeastern US While studies have begun to evaluate mastication as a fuels treatment option, much of this research has been in the western US and in ecosystem s where post treatment fuels are primarily woody dominated surface fuelbeds. Mastication ("mowing") in palmetto/gallberry pine flatwoods results in unique surface fuelbeds dominated by litter While surface fuel loading precisely controlled fire behavior during small scale fire behavior experiments, recovering shrubs controlled fire behavior in field scale experiments in these treatments. Fuel models have been a common approach to categorizing fuels for fire behavior prediction. Developing fuel models f or mastication treatments will need to take into account the ability of shrubs to resprout following these treatments. The fast recovery of shrubs following mastication in these flatwoods sites, along with their control over fire behavior, suggest that a shrub model would be appropriate for these treatments as soon as six months treatment. Unless sites are burned right after treatments, shrubs will dominate fire behavior. Treatments were effective at reduc ing fire behavior by reducing shrub biomass however longev ity of this treatment may be short lived as shrubs recover rapidly. Moreover, while shrubs control fire behavior, long duration heating from combustion of surface fuels may influence fire effects. Surface heating, observed during small sca le experimental burning, may have contributed to tree mortality observed during summer season burns conducted in the field. These flatwoods sites are highly flammable and
176 have likely adapted to fast burning shrub fires with significant intensity. Althou gh these southeastern pines are very resilient to crown damage ensued from burning, they are more susceptible to fine root and basal cambium damage when surface fuels burn for long durations. Mastication, while only reducing shrub biomass in the short ter m, increases surface fuels. Since treatments are likely to be prioritized in long unburned stands where duff has accumulated, adding surface fuels may result in increased ignitability of duff and potential overstory mortality. Burning in drier conditions to increase surface fuel consumption, a likely objective during prescribed burning in masticated stands, could pose a hazard to overstory trees if duff moisture is also low. Burning when surface fuels created from mastication are dry enough for consumpti on, but when duff is moist enough to limit damage to trees may be key to successful fuels management using these treatment regimes. Bulk density increases observed following treatments, immediately and one year following, may mean that fuels will be even more difficult to consume as time since treatment increases. This, along with shrub recovery, both indicate that follow up burning in these treatments should be conducted early to sufficiently reduce fuel loading and increase fire control. Developing tre atment regimes so that treatment timing will enhance meeting management objectives will be important. Mastication had minor effect s on the ecological attributes assessed with this research. Vegetation communities were little affected by treatments, exce pt that saw palmetto reduction was evidenced. S hrubs that vigorously sprout following burning may resprout following mastication because meristematic tissues and underground carbohydrate reserves are not destroyed. Apical m eristems in saw palmetto howev er, are embedded in the above ground stem and while they are typically not damaged
177 during burning, thus continuing to produce new fronds, they may be damaged by masticators during treatments. Understory or groundcover vegetation communities may change ove r time with a loss of palmetto cover, however only little evidence of increases in grass cover were observed here. Continued monitoring may reveal potential changes. Impacts of treatments on microclimate was minor, but treatment influences over fuel mois ture indicated that loss of shrub cover may have enhanced drying of surface fuels. While increased fuel bulk density should provide a mulching effect, drier surface fuels in masticated sites may actually increase ignition probability. Consumption of surf ace fuels may be aided by such an effect, however the risk of wildfire could be also enhanced. Moisture content in living shrub foliage wasn't influenced by mastication alone, however burning i n masticated sites resulted in shrubs with higher moisture con tent compared to sites burned that had not been previously ma sticated. Differences were likely due to reduced shrub cover, especially saw palmetto, and less competition for resources. Whether mastication is conducted as a stand alone treatment or follow ed up by prescribed burning, palmetto/gallberry pine flatwoods seem to recover quickly following treatments. Treatment effectiveness is likely not to last long without follow up burning. While concerns regarding potential impacts to overstory trees durin g burning in these treatments will need to be considered it appears that such treatments will likely have minor ecological impacts if conducted in a manner to minimize potential long duration surface heating. Considerations regarding treatment timing and conditions for follow up burning will need to be taken into account to minimize such impacts and meet management objectives. Palmetto/gallberry pine flatwoods are unique in their post
178 mastication fuel environment and provide additional insight into the effects and efficacy of mastication treatments as a whole.
179 LIST OF REFERENCES Abrahamson WG (1984 a ) Post fire recovery of Florida lake wales ridge vegetation. American Journal of Botany 71, 9 21. Abrahamson WG (1984 b ) Species re sponse to fire on the Florida lake wales ridge. American Journal of Botany 71, 35 43. Abrahamson WG, Hartnett DC (1990) Pine Flatwoods and High Prairies In 03 149 (Univ. of Central Florida Press, Orlan da, FL) Agee JK Skinner C N (2005) Basic principles of forest f uel reduction treatments. For est Ecol ogy and Manage ment 211 83 96. doi:10.1016/j.foreco.2005.01.034. Andrews PL, Bevins C, Carlton D, Dolack M (2008 ) Behave Plus Fi re Modeling System Version 4.0 USDA Forest Service, Rocky Mountain Research Station in cooperation with Systems for Environmental Management (Missoula, MT). Battaglia MA, Rocca ME, Rhoades CC, Ryan MG (2010) Surface fuel loadings within mulching treatments in Colorado coniferous f orests. Forest Ecology and Management 260, 1557 1566. Bradley T, Gibson J, Bunn W (2006) Fire severity and intensity during spring burning in How to Measure Success: Conference Proceedings USDA Forest Service, Rocky Mountain Research Station Proceedings RMRS P 41. pp. 419 428. (Fort Collins, CO) Brockway DG, Outcalt KW, Estes BL, Rummer RB (2010) Vegetation response to midstorey mulching and prescribed bur ning for wildfire hazard reduction and longleaf pine (Pinus palustris Mill.) ecoystem restoration. Forestry 82, 299 314. Brose P, Wade D (2002) Potential fire behavior in pine flatwood forests following three different fuel reduction techniques. Forest Eco logy and Management 163, 71 84. Brown JK (1971) A planar intersect method for sampling fuel volume and surface area. Forest Science 17 96 102. Busse MD, Hubbert K, Fiddler G, Shestak C, Powers R (2005) Lethal soil temperatures during burning of masticated forest residues. International Journal of Wildland Fire 14 1 10. Byram GM (1959) Combustion of Forest Fuels. In 'Forest fire: control and use'. pp. 61 89. (McGraw Hill: New York, NY)
180 Castro J, Allen CD, Molina Morales M, Maranon Jimenez S, Sanchez Mira nda A, Zamora R (2010) Salvage logging versus the use of burnt wood as a nurse object to promote post fire tree seedling establishment. Restoration Ecology 19 no. doi: 10.1111/j.1526 100x .2009.00619.x Catchpole EA, Catchpole WR, Rothermel RC (1993) Fire be havior experiments in mixed fuel complexes. International Journal of Wildland Fire 3, 45 57. Chapin FS, Matson PA, Mooney HA (2002) Principles of Terrestrial Ecosystem Ecology. (Springer Verlag: New York) Davis LS, Cooper RW (1963) How prescribed burning a ffects wildfire occurrence. Journal of Forestry 61, 915 917. Enriquez S, Duarte CM, Sand Jensen (1993) Patterns in decomposition rates among photosynthetic organisms: the importance of detritus C:N:P content. Oecologia 94, 457 471. Gagnon PR, Passmore HA, Platt WJ, Myers JA, Paine CET, Harms KE (2010) Does pyrogenicity protect burning plants? Ecology 91, 3481 3486. Glitzenstein JS, Streng DR, Wade DD (2003) Fire frequency effects on longleaf pine (Pinus palustris Mill.) vegetation in South Carolina and nort heast Florida, USA. Natural Areas Journal 23, 22 37. Glitzenstein JS, Streng DR, Achtemeier GL, Naeher LP, Wade DD (2006) Fuels and fire behavior in chipped and unchipped plots: implications for land management near the wildland/urban interface. Forest Eco logy and Management 236 18 29. Gholz HL, Perry CS, Cropper WP, Hendry LC (1985) Litterfall, decomposition, and nitrogen and phosphorus dynamics in a chronosequence of slash pine (Pinus elliottii) plantations. Forest Science 31, 463 478. Hood S, Wu R (2006 ) Estimating fuel bed loadings in masticated areas. In 'Fuels Management How to Measure Success: Conference Proceedings', 28 30 March 2006, Portland, OR. (Eds PL Andrews, BW Butler) USDA Forest Service, Rocky Mountain Research Station, Proceedings RMRS P 41, pp. 333 340. (Fort Collins, CO) Hough WA, Albini FA (1978) Predicting fire behavior in palmetto gallberry fuel complexes. USDA Forest Service. Southeast Forest Experiment Station Research Paper SE 174. Asheville, NC. 44 pp. Johansen RW, Wade DD (1 987) Effects of crown scorch on survival and diameter growth of slash pines. Southern Journal of Applied Forestry 11, 180 184.
181 Kane JM, Varner JM, Knapp EE (2009) Novel fuelbed characteristics associated with mechanically mastication treatments in northern California and south western Oregon, USA. International Journal of Wildland Fire 18, 686 696. Kane JM, Varner JM (2010) Understory vegetation response to mechanical mastication and other fuels treatments in a ponderosa pine forest. Applied Vegetation Scie nce 13, 207 220. Knapp EE, Varner JM, Busse MD, Skinner CN, Shestak CJ (2011 ) Behaviour and effects of prescribed fire in masticated fuelbeds. International Journal of Wildland Fire 20, 932 945. Kobziar LK, Stephens SL (2006) The effects of fuels treatment s on soil carbon respiration in a Sierra Nevada pine plantation. Agricultural and Forest Meteorology 141, 161 178. Kobziar LK, McBride JR, Stephens SL (2009) The efficacy of fire and fuels reduction treatments in a Sierra Nevada pine plantation. Internatio nal Journal of Wildland Fire 18 791 801. Kreye J K, Varner JM, Knapp EE (2011 ) Effects of particle fracturing and moisture content on fire behavio u r in masticated fuelbeds burned in a laboratory. International Journal of Wildland Fire 20, 308 317. Kreye J K, Varner JM, Knapp EE (2012 ) Moisture desorption in mechanically masticated fuels: effects of particle fracturing and fuelbed compaction. International Journal of Wildland Fire doi: 10.1071/WF11077 Lewis CE, Hart RH (1972) Some herbage responses to fire on pinewiregrass range. Journal of Range Management 25, 209 213. McNab WH, Edwards MB (1978) Estimating fuel weights in slash pine palmetto stands. Forest Science 24, 345 358. Menges ES, Gordon DR (2010) Should mechanical treatments and herbicides be used as fire surrogates to manage Florida's uplands? A review. Florida Scientist 73, 147 174. Moghaddas EEY, Stephens SL (2007) Thinning, burning, and thin burn fuel treatment effects on soil properties in a Sierra Nevada mixed conifer forest. Forest Ecology a nd Management 250, 156 166. Molina DM, Galan M, Fababu DD, Garcia D, Mora JB (2009) Prescribed fire use for cost effective fuel management in Spain. In: Proceedings of the Third International Symposium on Fire Economics, Planning, and Policy: Common Proble ms and Approaches, USDA Forest Service Pacific Southwest Research Station General Technical Report PSW GTR 227 pp. 370 374.
1 82 Neary DG, Ryan KC, DeBano LF (Eds) (2005) Wildland fire in ecosystems: effec ts of fire on soil and water. USDA Forest Service, Rocky Mountain Research Station, General Technical Report RMRS GTR 42 Volume 4. (Ogden, UT) Nelson RM Jr., Adkins CW (1986) Flame characteristics of wind driven surface fires. Canadian Journal of Forest R esearch 16, 1293 1300. O'Brien JJ, Hiers JK, Mitchell RJ, Varner JM, Mordecai K (2010 a ) Acute physiological stress and mortality following fire in a long unburned longleaf pine ecoystem. Fire Ecology 6, 1 12. O'Brien JJ, Mordecai KA, Wolcott L, Snyder J, Outcalt K (2010b) Fire managers field guide: hazardous fuels management in subtropical pine flatwoods and tropical pine rocklands. Final Report to the Joint Fire Science Program, Final Report JFSP 05 S 02. Pitman WD (1993) Evaluation of saw palmetto for bi omass potential. Bioresource technology 43, 103 106. Prescott CE, Blevins LL, Staley C (2004) Litter decomposition in British Columbia forests: controlling factors and influences of forestry activities. BC Journal of Ecoystems and Management 5, 44 57. Rhoa des CC, Battaglia MA, Rocca ME, Ryan MG (2012) Short and medium term effects of fuel reduction mulch treatments on soil nitrogen availability in Colorado conifer forests. Forest Ecology and Management 276, 231 238. Ross MR, Castle SC, Barger NN (2012) Eff ects of fuels reductions on plant communities and soils in a Pinyon juniper woodland. Journal of Arid Environments 79, 84 92. Rothermel RC (1972) A mathematical model for predicting fire spread in wildland fuels. USDA Forest Service, Intermountain Forest a nd Range Experiment Station Research Paper INT 115. (Ogden, UT) Rothermel RC, Deeming JE (1980) Measuring and interpreting fire behavior for correlation with fire effects. USDA Forest Service Intermountain Forest and Range Station General Technical Report INT 93. ( Ogden, UT ) Schwilk DW, Keeley JE, Knapp EE, McIver J, Bailey JD, Fettig CJ, Fielder CE, Harrod RJ, Moghaddas JJ, Outcalt KW, Skinner CN, Stephens SL, Waldrop TA, Yaussy DA, Youngblood A (2009) The national Fire and Fire Surrogate study: effects of fuel reduction methods on forest vegetation structure and fuels. Ecological Applications 19, 285 304.
183 Scott JH, Burgan RE (2005) Standard fire behavior fuel models: a comprehensive set for use with Rothermel's surface fire spread model. USDA Forest Ser vice Rocky Mountain Research Station General Technical Report RMRS GTR 153. (Fort Collins, CO) Stephens SL, Moghaddas JJ, Edminster C, Fiedler CE, Haase S, Harrington M, Keeley JE, Knapp EE, Mclver JD, Metlen K, Skinner CN, Youngblood A (2009) Fire treatme nt effects on vegetation structure, fuels, and potential fire severity in western U.S. forests. Ecological Applications 19, 305 320. Tanner G, Mullahey JJ, Maehr D (1996) Saw palmetto: an ecologically and economically important native palm. IFAS Circular W EC 109, University of Florida, 3 p. (Gainesville, FL) Vaillant NM, Fites Kaufman J, Reiner AL, Noonan Wright EK, Dailey SN (2009) Effect of fuel treatments on fuels and potential fire behavior in California, USA, national forests. Fire Ecology 5, 14 29. V an Lear DH, Carroll WD, Kapeluck PR, Johnson R (2005) History and restoration of the longleaf pine grassland ecosystem: Implications for species at risk. Forest Ecology and Management 211, 150 165. Van Wagner CE (1973) Height of crown scorch in forest fire s. Canadian Journal of Forest Research 3, 373 378. Varner JM, Kush JS, Meldahl RS (2000) Ecological restoration of an old growth longleaf pine stand utilizing prescribed fire. In 'Fire and forest ecology: innovative silviculture and vegetation management: Tall Timbers Fire Ecology Conference Proceedings Tallahassee, FL (eds K Moser, C Moser) Tall Timbers Research Station Conference Proceedings 21, 216 219 (Tallahassee, FL) Varner JM, Gordon DR, Putz FE, Hiers JK (2005) Restoring fire to long unburned Pi nus palustris ecosystems: novel fire effects and consequences for long unburned ecosystems. Restoration Ecology 13, 536 544. Varner JM, Hiers JK, Ottmar RD, Gordon DR, Putz FE, Wade DD (2007) Overstory tree mortality resulting from reintroducing fire to l ong unburned pine forests: the importance of duff moisture. Canadian Journal of Forest Research 37, 1349 1358. Varner JM, Keyes CR (2009) Fuels treatments and fire models: errors and corrections. Fire Management Today 69, 47 50. Waldrop TA, Van Lear DH ( 19 84 ) Effect of crown scorch on survival and growth of young loblolly pine. Southern Journal of Applied Forestry 8, 35 40. Westerling AL, Hidalgo HG, Cayan DR, Swetnum TW (2006) Warming and earlier spring increases western U.S. forest wildfire activity. Scie nce 313, 940 943.
184 Zipperer W, Long A, Hnton B, Maranghides A, Mell W (2007) Mulch flammability. In Emerging Issues Along Urban Rural Interfaces II : Linking Land Use Science and Society: Edited by DN Laband ) pp. 192 195
185 BIOG RAPHICAL SKETCH Jesse Kreye was born in 1974 in Mora, Minnesota and subsequently raised in the North Star state. He graduated from Hinckley Finlayson High School in 1993. He served in the U.S. Navy as an aircraft firefighter aboard the U.S.S. Enterprise. Jesse received a Bachelor of Science in forestry in 2005 and a Master of Science in natural resources with a minor in biometrics, in 2008 from Humboldt State University. He received his Ph.D. in forest resources and conservation in 2012 from the Univer sity of Florida. He has worked for the Minnesota Department of Natural Resources in fire suppression and the U.S.D.A Forest Service in fire management silviculture, a nd wildlife management. Jesse was also a lecturer in the Department of Forestry and Wil dland R esources at Humboldt State University