Group Title: impact of wet season and dry season prescribed fires on Miami Rock Ridge Pineland, South Florida
Title: The impact of wet season and dry season prescribed fires on Miami Rock Ridge Pineland, South Florida
Full Citation
Permanent Link:
 Material Information
Title: The impact of wet season and dry season prescribed fires on Miami Rock Ridge Pineland, South Florida
Alternate Title: Miami Rock Ridge Pineland, South Florida
Physical Description: viii, 154 leaves : ill. ; 28 cm.
Language: English
Creator: Snyder, James R
Copyright Date: 1984
Subject: Prescribed burning -- Florida   ( lcsh )
Forest fires -- Environmental aspects -- Florida   ( lcsh )
Pine -- Ecology -- Florida   ( lcsh )
Botany thesis Ph. D
Dissertations, Academic -- Botany -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Statement of Responsibility: by James R. Snyder.
Thesis: Thesis (Ph. D.)--University of Florida, 1984.
Bibliography: Bibliography: leaves 110-122.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00099348
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000471635
oclc - 11898805
notis - ACN6476


This item has the following downloads:

impactofwetseaso00snyd ( PDF )

Full Text








Tha major portion of this work was supported directly by

the South Florida Research Center, Everglades %ltional Park,

and I owe s3ezial thanks to Dalq Taylor, formerly fire

ecologist, ai Gary Hendrix, research director. Numerous

individuals it th= Pesearch Center aided in the filll wor:;

in particular Rebecca Rutlelge, Virginia Louv, Arthony

Canrio, Laeis Sharman, ani Donna Blake spont many hours

under trying conditions. Alan Heraion, besides helning with

fill work, identified plants and shared his knowledae of

local plant :ommanities. Pill Robertson, Jr., hloe1 snrac

my interest in tasse unusual pinlands. The ?or? r=snur7?

manaqgemnt staff (lead by fire bosses Phil Foeon, Leon '~nz,

ini Ron Sutton) 3arriel out the prescribed burns. ?aul rrt!

of the AgrLcultural Research and Education Center,

Homestead, generously permitted the use of his facilities

for tissue grinding.

In gainesville my committee chairman, Jack Ewel, provided

logistical sioport, advice, and guidance. Committee nembbrs

Seorge Bowes, Walter Judd, and Tugh Popenoe made useful

comments and suggestions throughout the study. Jack Pu'z

provided a tiorogh review of the manuscriot and sqsested

numerous impcovenents. Mary McLeod, Craig Peed, and Linda

Lee of the Forest Soils Lab were of great technical

assistance with the chemical analyses. George u1ller

drafted most of the figures. Xen Portier aevised on

statistical nattscs. computing was done with the facilities

of the Northeast Regional Data Center.

In recognition of less tangible contributions, I thank n7

father, Robert Sayder, for instilling in ae an aopre-iation

of the great out-of-doors and my wife, Jean Snyder, for

continuing encoucagement and support.


ACK O L G TS . . . . . . . . . ii

ABSTS ACT . .. . . . . . vi


I. NTRODU TCrT . . . . . 1

Dbjpctivos . . . . . . . .. . .
he ecosystem . . . . . . 3
Cli e . . . . . . . . 3
a-ol3gy and Soils . . . . . . . 7
Vegetation . . . . . . . 11
'ire Ecology . . . . 17
?rPasnt Extent an- Conlition . . . .
7verVlades National ark . . . . 2

II. 'T.O S . . . . . . . . . . 27

7xporimnatal Design and Site Selection . 27
The Burns . . . . . 29
Abajeqrannd Mass Sampling .. . . . . 31
eagetatian and Litter . . . , . 31
?3stburn Ash . . . . . . . . 34
Litterfall . . . . .. . . . 35
Soil Sampling . . . . . . . .. 36
issue and Soil Analysis . . . ... . 37

ITT. RESULTS . . . . . . . . . . . 39

Tnitial ageattion Structure . . . . .. 39
Burn Descriptions . . . ... . .. . 4
Pica Effects an Unlerstory Mass aan Nutrients .r2
faitial Distribution of Mass an! Nutrients 52
ksh Collection methods . . . . .
?ost3urn Distribution of Mass ard Nutrients 57
Postfica Recovery of Mass and Nutrients . 58
vegetation .9.. .. ... .... . B
Littr. ...... .......... .. .70
Impact of Fires on Pines . . ..... .. 75
Soi . . . . . . . . . . .. . 79

IV. DTSC SS N . . . . . . . . . 32

Fice-caused Losses of Nutrients . . ... 2
?3stfice Recoverv . . . . . . 0
TfE~at 3f Season aE Burning on Hardwool recovery 93
Regrovti Vegetation as a Nutrient Sin .- . 101
Sumnary and Conclusions . .. . . 10

LITE RAT ED . . . . . . . . . 11




1ETB PAPHICAL SETCrH . . . . . . . . .

Abstract of Dissertation Presonte~ to the (raduite School
of the Universitv of 7lorida in Partial F lfillmeon' f the
Requir3en2ts for the Degree of Doctor of Philosophy



James R. Snvyer

August 1984

Chairman: Join J. Ewel
Major Department: Botany

In the subtropical pine forests on oclitic li.mstone in

Dade Countv, "lorida, "inu3 elliottii var. 3easa qrows over

a species-cizh unaarstory (> 128 sop.) 3f shrabbh hardwoods

(mostly tropi-cl evergreen species palms, and herbs,

including several endemic taxa. Fires prevent rapid

conversion to hardwood forest. To compare the response of

these pinelais to burning during the lightning Cire season

(hot, wet, summer months) and the management fire season

(cooler, drier, winter months), aired burns were condnc't?

at two sites in Rverglades national Park, one bi=rnd 3., vr

previously and the other 5 yr. Aboveground inderstory

biomass and attriants were measured immediately before and

after, and it 2, 7, and 12 mo after the four burns.

The burns topkilled all the understory vegetation. Thp

fires volatiLizel 1-1.5 kg/m2 of organic matter and 5.7-".

g/m2 of N. Meteorological inputs and symbiotic and

nonsymbioti Efixation should easily replace N lost in the

burns. Losses of P, K, :a, and Ig were not letectable

except for K in one of the burns.

The rapil un]irstcry recovery was almost entirely

vegetative rgrowth of the topkille9 plants. 'ine seenlin4

were abundant after the wet season burns, however. fera;

and palms recovered dry mass more rapidly than hardwoods an,

reached prebucn levels within 1 yr. Hardwoods recovered

only, 18-39% of their preburn biomass. Total understory

vegetation re3ovecy was 27-63: of initial amounts, but

leaves rpecovceI 58-931. Net primary prodlctivitv t'- first

year after burning was 144-200 g/m2.

Recovery if nutrients was more rapid than bioass because

of higher nutrient consentcations in regrowth tissues. Some

herb and pall nutrient standing crops reached preburn levels

within 2 mo.

After the burns litter mass and nutrients often showed an

initial decrse before recovery began, kt 1 vr litter iass

was 42-62" of the prebnrn amount. Annual pinon no1=lef-ll

averaged 260 and 320 g/m2 at the two sites,

The iaouat ff hacrwood recaverv was not determined by

season of burning; higher fire temperatures (wet season b'rn

in one case, 3ry season burn in the other) resulted in less





The importance of fire as an environmental factor

influencing ecosystem structure and function is widelv

appreciated today (Ahlgren and Ahlgran 1960, KozlowsIi and

Ahlgran 1974, Moanev et al. 1981, Rundel 1981, Wright ana

Bailey 1982). Fire is often thought of as a succ-ssion-

initiating listurbance (White 1979). However, in

communities that iave evolved under a regime of frequert

fires, it is tie exclusion of fire that may result in

3raiatic--if not su3ian--changes. Much of sonth?-stern

U.S., particularly the coastal plain, is covere wvith

vagatation that requires periodic burning (Chris~'nsn

1981). lany southern pine forests, for example, levelon

into hardwool forests in the absence of burning warren n


The regrowth vegetation in some frequently burned '

ecosystems comes Erom seeds present in the soil or release!

from killed plants. gore commonly, however, individuals

survive and sprout back from belowgrouni parts. In

chaparral it is common for species to show mixed s'-1ling

and sproutinJ recovery mechanisms (Keeley and Koeley 1991).

Fire is very prominent in South Florida ecosystems, both

in terms of area burned and as a leterminant of vegetation

pattern (,obertson 1953, Wa3e et al. 1980). Of particular

interest in this study are the subtropical pinelands founi

on the Milni Rock Riige in southeastern vlorida. These Dine

forests differ Erom other southeastern coastal plain oina

forests in two ways: they include a large number of

tropical species in the understory and they grow directly on

a limestone substrate nearly devoil of soil.

Prescrib.l fire is used throughout the southern nine

region for site preparation, fuel reduction, and range

improvement. It has also become a tool in natural area

management in places such as Everglades National ?ark. Tn

most southeastern pine forest types (with one maior

exception) fire loes not kill the canopy trees as it locs in

many conifernos forests (Heinselman 1973). This, along wilt

its common use for other purposes, may be why the us0 of

intentionally set fires was so readily accepted as

management techniLue in natural areas.

Prescribed burns in the southeast are traditionally

carried out lurina the cooler months because fires durirg

this time are less likely to damage overstory trees and

because the casprouting vegetation provides forage vhich is

otherwise limited at this time. Before the influence of

humans, however, fires were primarily ignited bv lightning,

which occurs during warmer months.


i wished to examine some ecosystem-level responses on

Miami Rock Ridge pinelands to fires during the natural

lightning-cuaisd fire season and the traditional nanagament

fire season. T was particularly interested in the


(1) The amount of mineralization and loss of organic

matter and natrients due to fires.

(2) The pattern of recovery of understory mass and

nutrients during the first year after burning.

(3) The degcre of recovery of the unnarstorv 1 yr af'er

burning, viti :mplasis on the recovery of hardwoods.

(4) Rasad on the above items, to draw some conclusions

about the potential role of fire in nutrient cycling and the

use of pres=ribeS fire in the management of a natural area.

The ecosystemm


The cliiate of South Florida has two salient features:

(1) moderate, almost frost-free winter t=aperatures and f2)

a marked seasonality in rainfall (?ig. 1). These

characteristics result in a hot period of high rainfall

(May-Oct.) and a cooler period of much lower precipitation

(Nov.-April These are known as the wet and Irv seasons,

respectively, in keeping with tropical terminology.






15- -


.555 F;~ :. .... ..... ,.. .

Figure 1. Mean monthly precipitation (bars) and
temperature (dots) at Royal Palm Ranger Station,
Everglades National Park (after Rose et al. 1981). Mean
annual rainfall is 146.3 cm.

Mean montily temperatures range from about 18.0oC in

December an3 January to about 27.50C in July and August.

Prosts occur in the homestead area about once every two

years on the average (Brallev 1975). These frosts can

damage winter vegetable crops and production of avocados and

mangos for the ensuing year. Many of th3 native plant

species are also susceptible to Erost damage (Craigheal

1971), soaecially plants in open areas. freezingg

temperatures were recorded during both 1981 and 1982, bit

the study areas were little affected. Only a few of t'h

minor species dropped leaves and none hal stems kili"d.

Total annual rainfall averages 146 cm in the southern

portion of tia Miami Rock Ridge, with almost 90: (117 ca)

coming during the six wet season months (Fig. 1). 7ater

levels vary seasonally with a maximum in Senotaber and a

minimum in Aoril, During the wet season clous" bnil, un iL

the afternoon and result in brief thundershowers; the

lightning that accompanies these storms is a potential

ignition source for wildfires. Although most summer

rainfall is :onve:tional, tropical cyclones can bring large

amounts of rain. In August 1981 Long Pine Key in Everglades

National Pinr received more than 40 cm of rain in thrs dItys

from tropical storm Dennis.

The soutaeastern coastal area of 'lorida can exact a

tropical cyclone once every 5 yr and a hurricane-fnrce storm

once everv 7-9 yr (Gentry 1974). The damage to vegetation

in South Fladria by hurricane Donna in 1960 was substantial,

especially in th~ mangroves on the southwest coast

(Craighead and gilbert 1962). The pinelands of Wveralades

National Park suffrsed little damage in spite of

experiencing wind speeds greater than 160 km/hr. Th? hiah

water levels Orought on by heavy precipitation can affect

the pineland vegetation more strongly than the high vinds.

although South Florida is north of the Tropic of Cancer

(Long Pine Key is about 250 23' N latitude it is commonly

referred to as "tropical Florida," especially hv those

concerned with fl3ristics (e.g., Tomlinson 19i0, Lona and

Lakela 1971). In fact, a world-wiie climatic classification

scheme based on that of Koeppen (Ccitchfield 197u) considers

the southern tip of Florida to have an Aw, or tropical

savanna, linate of the vet-and-dry tropics. It is included

as a tropical climate only because the mean monthly

temperature of the coolest month is greater than 180C. T"-

common occurrence of frost at sea level would perhaps ma~k

subtropical a better designation for the climate. The

classification system of Holdridge (1947), which is based on

temperature ian precipitation, places southern Florida in

the Subtropical Ioist Forest life zone.

Geolo yand Soils

South Florida is extremely flat: a function of its

marine depositional history, low elevation, and relativelv

short periol of emergence. The broad, shallow Pverqlades

basin which ertenis south from Lake Okeechobee is bounded on

the east by the slightly higher Atlantic coastal ridge. Thp

southern end of this ridge in Dade County is an outcropping

of oolitic limestone known as the !iami Rock Ridge (Davis

1943). This region, previously dominated by pine forests,

extends from the vicinity of ?iami southwestward to

Homestead ani westward into Everilades National Park (Vin.

2). The mai[mum elevation of the ridge is about 7 m in

Coconut Grove (Hoffmeister et al. 1967) and it drops to less

than 2 m in everglades National Park, where it disappears

under the surrounding wetlands,

The iiami Limestone (rqoffmaister at al. 1967) which m'ake

up the Miami Rock ?idje is the surface rock of virtulallv all

of Dade County. It represents Pleistocene marine deposition

of calcium carbonate during the Sangamon stage (Cooke 19ac).

The upper oolitic faces which forms the rock ridge is

composed of Doids, pellets, and some skeletal sand. To th=

north along the coastal ridge the limestone is blanketed by

a layer of Pamlico sand, and at lower elevations to the west

and south it may be covered with late Pleistocene or "eccen

marls and pets. Several transverse depressions nassina

through the ridge represent valleys in tha rock that have

been partially filled by deposition of marl and/or organic

matter (igs,. 2 and 4).

The formation of the Miami Rock Bidge is described hv

Hoffmaister :t al., (1967). They compare it to processes

occurring tolay on the northwest section of the greatt Bahama

Bank, where loose mounds of ooids are forming and shifting

in the shallow water on the eastern edge of the Straits of

Florida. The tilal channels that cut through the broal

ridge of unconsolidated oolitic sediment are thought to be

analogous to thos that form the transverse valleys in 1h,

Miami Sock Ridge.

The oolite rock is soft and friable until indurated hv

exposure to the atmosphere. Dissolution of the rock has

left the surface honevcoibed with numerous holes and

fissures, anl armed with sharp, jagged projections. In i's

most treachercus form it is known as pinnacle rock. Th'

diameter of solution holes can range from centimeters to

meters as can the depth, although 0.5 m diameter and 1 m

depth might be common dimensions for the larger holes. The

character of the rock surface varies from place to olace,

with differences in the degree of solution and the amount of

loose rock fragments on the surface.

In the pineland areas of the Miami Rock Pilqe the scanty

soil is founi in solution holes, depressions, and cracks in

the rock. The soils are members of the Rockdale series,

which is classified as a Lithic Ruptic-Alfic -utrochrcpt,


clayey, mizel, hyperthermic (Soil Survey Staff 1975), 'he

surface soils range from dark grayish brown to brown fine

sands or fin* sandy loams. The subsurface layers (where

present) are light gray to yollowish-red fine sand and brown

to reddish-bcown sandy or clay loams (Soil Conservation

Service 1958). In the northern portions of the ridge in the

Miaai area, the soils tend to be very sandy due to

deposition oF sands during the Pamlico stage. To the south

there are often reddish-brown residual soils; the

predominance of these soils in the area !ist north of

Homestead is responsible for the appellation "3llands" qivan

to the area. In most areas there is little soil ezposed on

the surface ind plant roots run through cracks and channels

in the rock that are filled with a mixture of organic matter

and weathering products of the limestone. causee the roc'

ridge is the highest part of 'he landscape and is formc of

porous limestone the soils are generally well-drained,

However, high water tables can reach the ground surface in

the lower lying areas during the wet season. The soils are

neutral or slightly alkaline in reaction and are deficient

in N, P, and K for agricultural crops. The fact that the

exploitable soil volume is so small also contributes -n

making this i very oligotrophic ecosystem.

The nearly frost-free dry season makes southern vllrila

in important production area for winter vegetables. Current

farming practices require extensive site preparation before


production. After bulldozing off the native vegetation the

surface 15 cm of rock is scarified with heavy eguinment

(rock plowes), resulting in a "soil" composed of a mixtur-

of rock fragnants, pulverized limestone, and a saill iaount

of original soil. This treatment increases the rooting

volume and raises the pa t3 8 or higher. Crops (such as

tomatoes, squash, and pole beans) req'iire heavy

fertilization and irrigation. Experience in Everglades

National Parr suggests that once the substrate has bee-

altered in tais way the native vegetation does not normally



The vngetition of the Miami Rock Ridge is essantiallv a

mosaic of three basic vegetation types: pineland, hamoc',

ani prairie. Te pinelands form the dominant matrix on -he

higher grounI, with small islands of tropical hardwood

forest known as hammocks scattered within this matrix. Tn

the shallow transverse depressions that run through the

ridge are herbaceous prairies, or glades, similar to the

vegetation bordering the ridge. Long (1974) estimates that

the vegetation of South Florida is about 5,000 yr old.

2nsnral dessciptions of the natural vegetation of southern

Florida, including the rock ridge, have been provided by

several authors: Harshberger (1914), Harper (1927), Davis

(1943), Craijhead (1971), and Rade et al. (1990).

hammocks ice patches of closed forest of essentially

evergreen, bcoadleafed trees. They are found in the higher

areas of the ridge; where they are seldom flooded, if ever.

The soil is a thick layer of organic matter that has

accumulated on the surface of the limestone. Tmoortant

canopy species include Nectandra coriacea, Coccoloba

diversifolia, 212E c S IEiriniana, Lysiloma latisi5ionm,

12Metpiu toiferau, Ficus aurea, gBumlia salicifolia, and

Bursera simacuba (Phillips 1940, Alexander 1967, Craighea!

1974, Olmste3 et al. 1980a). Of these only Q, virinia.ns is

not a typically tropical species. The unierstory consists

of saplings of the canopy species, a number of small tree

species, and a a few shrubs. Relatively few herbs are

present on the ground, but epiphytic bromeliads an! -rchis

are quite common.

The seasonally flooded prairies that border the lovwr

fringes of tie pinaland are Dominated bv grasses and sedes

(e.g. juhlenberjia filip2s and Clalium jnamaicens:) but

include numaeous forbs (Pocter 1967, Olmsted et al. 1983).

Here the water table is above the ground surface for 2-'1

3m/yr and the limestone is topped with a thin layer of marl

(Olasted at al. 1980b).

The pinelinds, which oc-upy most of the iami pock Ridge,

are monospecLfic stands of Pinus elliottii var. tn4a, the

South Plorida slash pine, with a diverse understory of

palms, hardwoods, and herbs. The South Florida varietv o.


slash pine differs from the commercially important northern

variety (var. alliottii) in several respects (Little and

Dorman 1954a,b). lost conspicuous is the grass-like

seedling stige reminiscent of Pinus palustris (lonqleaf

pine) in which the seedling grows for several years without

stem slongatLon. wavever, Squillace (1965) showed more or

less continuous variation across the range of slash pine for

12 traits, i2cluiing seedling height and stem diameter. "'ho

range of Souti Florida slash pine extends from the Lower

Florida Keys to about Polk County in the center of

peninsular Florila and Levy and Volusia Counties alonq the

Gulf and Atlantic coasts, respectively (Langdon 19A3).

Sandy flatwools are a more widespread habitat for South

Florida slash pine than the limestone of the "iami lock


In contrast to the monotony of the pine overstory, +h-

understory is relatively species-rich. The shrub layer is

composed of 15-25 hardwood species per 0.16 ha (Loope et al,

1979), most of which are tree species maintained as shrubs

by repeated fires; the palas Sabal palmetto, Serenoa reens,

and Coccothrinax ar entata; and a cycad, Zamia numila.

Hardwoods commonly found as shrubs in the pinelands and as

trees in himmocks include letonium toxifarum, Sum-lia

salicifolia, srXEsine flori3ana, Guettarla scabra, anC

Ardisia escallonioides. Most hammock tree snpcies are founI

at least occasionally growing in pineland. Some smaller

shrubs (e.g. Dolonaea viscosa, Lantana de rEEssa, and t.

involucrata) are found in the open oinelanas but not in the

shade of hammocks.

Most of the hardwoods are West Indian in distribution and

are found only in extreme southern Florida or along the

coast to northern Florida. Only a few of the more important

species are founl as far north as Gainesville, 'lorida:

Rhus conalli a, zrEica cerifera, llex, Persea

borbonia, ini Q!erzus virriniana. Notably absent from th-

rock rid3j pinalinds are llsx lab hp (gallberrv) and members

of the Ericaceae, s3 important in most southeastern

pinelands. Apparently the high soil OH excludes these

species. Tie biogeography of pineland shrubs in South

Florida, including the Miami Rock Ridge, has been detailed

by Robertson in Olmsted et al. (1993).

The herb Layec is dominated by grasses but also contains

seiges, forbs, and three common ferns. The number of herh

species per 0.16 ha varies from 50 to 75 (Loope et al.

1979), The relative importance of hardwoods, palms, an1

herbs varies depending on local elevation and fire history.

In the lower, vwtter pinelands the understory tends to have

fewer hardwoods and has an herb layer that shares many

species with the prairies. ?reaaently burned sits have

better developed herbaceous layers than infrequently burned


Loope et il. (1979) list 196 plant taxa for the rock

ridge pinelaais, an3 67 of these are restrict!I to pineland

habitats. Tha number of South Florida endemics found in

pinalands (32, 17 of which are found exclusively in

pineland) is by far the highest for any South Florida

vegetation type (Avery and Loope 1980a). All the endemics

are herbs ex~:pt for the shrubs ForestiEra speqceata var.

2inatorum an Laitana L dl erssa (Table 1). South 'lorida

slash pine, although found north of Lake Ckeechobee, is

endemic to peninsular Florida and the Florida Keys.

The pinalians most similar to those of the Miami Pock

Pidge ire those of Pig Pine Key and several other Lover Fnvs

where the Miami Limestone also outcrops. These pinelands

differ mainly in the presence of several tropical hardwoods

characteristic of the Florida Keys or nearby coastal areas

and the oroninence of tree-size~ palms of Coccothrinav

arqgntata and Thrinax morrisii (Alexander and Dickson 1972).

The pinelanis of the Bahama and Caicos Islands are also vsry

similar to those of the rock ridge except that tha pine is

P. caribaea var. bahamensis (March 19U9, Luckhoff 1964,

Lamb 1973). The substrata is weathered coralline limestone

very much lit the oolitic rock of South Florida. many of

the understocy species are the same as those found in rock

ridge pinelinas (Coker 1935, Correll and Correll 198?),

although Robertson (1962) noted the conspicuous absence o"

Sersnoa renens. Even though it grows in an ecological


Table 1. Vascular plant taxa endemic to South Florida and
found in Miami Rock Ridge (MRR) pinelands,
including those of Everglades National Park
(after Avery and Loope 1980a).

Exclusively Present in
Taxon in MRR Everglades
Pinelands National Park

Amorpha crenulata *
Argythamnia blodgettii *
Aster concolor var. simulatus *
Brickellia mosieri *
Chamaesyce conferta *
C. deltoidea var. adhaerens *
C. deltoidea var. deltoidea *
C. garberi *
C. pinetorum *
C. porteriana var. porteriana *
Croton arenicola
C. glandulosa var. simpsonii *
Dyschoriste oblongifolia
var. angusta *
Evolvulus sericeus var. averyi *
Forestiera segregata *
var. pinetorum
Galactia pinetorum *
G. prostrata *
Hyptis alata var. stenophylla *
Jacquemontl curtisii *
Lantana depressa *
Linum arenicola
L. carter var. carter
L. carter var. smallii
Melanthera parvifolia *
Phyllanthus pentaphyllus
var. floridanus
Poinsettia pinetorum *
Polygala smallii *
Schizachyrium rhizomatum *
Stillingia sylvatica ssp. tenuis *
Tephrosia angustissima *
Tragia saxicola *
Tripsicum floridanum *

setting virtually identical to that of the roci ridge, .

caribaea vic. bahamensis do0s not have the grass-like

seedling stage that characterizes South ilorida slash pine.

It should be noted that for many years P. gll.l-tiii o

southeastern U.S. was considered the same species as P.

caribaea of the West Tndies and Central America P(Little ard

Dorman 1954a,b)

Fire Fcology

The rock cidge pineland, like most southern pine forests

(!arran 1943), is a fire-maintained vegetation type that

develops into hardwood forest in the absence of turning.

Robertson (1953) estimated that within 15-25 vr of the

cessation of burning, open pine forest would become sensee

hardwood focSst (hanaock) under a stand of relic pines; this

happened in a small area of pine forest in 7ver la=s

National Pack that was protected from fires by the

construction of a road (see photographs in Wade et al. 10)0,

p,93), Alaian~ic (1967) documented the rapid succ-ssion o'

pineland to hammock elsewhere in the rock ridge. The

succession to closed hardwood forest results in the

elimination of sose shrub species and the rich herbaceous

flora characteristic of pinelands (Loope and Dunevitz 1991).7

This is probably due to the reduction in light ceachinq the

ground but it may also ba due in part to the thick

accumulation of organi' matter that is normally removed by

fires. Loops and Danevitz (1981) found fewer species in a

frequently barnel piaeland than in a pineland unburned For

35 yr.

Fires in pinelands are surface fires that move along the

ground consuming litter and understory vegetation. Thp

density of the pine canopy is such that crown fires are

unknown, although the trees can be killed by convectional

heat under severe burning conditions. Prairies burn readily

when sufficient fuels are present but hammocks under normal

circumstan-cs do not. Pineland fires usually burn lin to th"

edge of hammaiks ani go out; however, during extreme

droughts the fire nay smolder through the hammock, consuiin'

the organic soil and killing the trees, Craighead (197U)

suggests that once soil moisture content in hammocks drons

to 351 they ire susceptible to soil fires.

All the species present in thq oinelanis are adapted '

fires. ~lture slash pines are very resistant to fire

because of a thick, insulating bark and the relatively heavy

buds surrounaed by long needles (Hare 1965b, Bvram 19U8).

Both varieties of slash pine are able to recover from 100'

crown scorz= in some cases (Wade 1983, personal

observation). Besides being fire-resistant as an adult, the

South Florida slash pine has a fire-resistant seedling stag

much like lonaleaf pine (Littl- and Dorman lq54a,b). An

accidental fire in a plantation of both var. densa and var,

elliottii seeilings killed a smaller proportion of th? South

Florida variety (Ketcham and Bethune 1963). Seedling

establishmint is also favored by fires occurring soon before

seed fall (Klukas 1973).

The hardwoods in the understory all have the ability to

sprout back after being topkilled. Some, such as Dhus

g oEllia, are prolific root-sprouters, but most sen! up new

shoots from the rootstock at the base of the stem.

generally, few shrubs are tilled completely by single fires.

Robertson (1953) recorded mortality of 0-10" in nine specie-

after a fire. Depending on the funl accumulation and

burning conditions, some fires may have no annarent Cfcact

on larger hardwoods. All the herbs arn perennials that

sprout back auizkly and seem to flower more profusely in

recently burad areas than in unburned areas. Some of t'is

apparent increase in reproductive activity may be lue to

improved visibility or more synchronous flowering, hut thcr

is no question that flowering of many species of grasses is

stimulated by burning (Robertson 1962).

The fire history of the rock ridge pinelands is difficult

to reconstruct with any degree of detail. The use of fire-

scarred trees is limited because of the small number of

suitable traes and the ambiguity involved in ring counts in

South Florida slash pine (Tomlinson and Craighead 1972, Arno

and Sneck 1977, Taylor 1980). The only methodology

available is to deduce from present-day fire patterns and

species attributes what might have occurred in the past.

Lightning fires are a common occurrence today and are

likely to have been so ever since the most recent emergence

of South Plorila from the sea. The lightning is pronuce-

during freqaunt convectional storms during the wet season.

The average annual number of lightning strikes reaching the

ground in the rock ridge region is 4-10 per km2 (Taylor

1980). The =ommon sight of single dead pine trees with

longitudinal fissures running down their bark is testimony

to the high incidence of lightning strikes. In Fverqlad-s

National Part lightning-caused fires accounted for 2P' o'

all fires ini 181 of the park area burned during the peri1-

1948-1979 (Taylor 1981). Almost all these fires occurr?!

within the wet season months of "ay to October (?ig. 3).

During this study (on Aug. 2, 1990) a small lightning fire

burned about 3 ha of pineland near one of the sites.

Egler (1952) assumed a low frequency of lightning fires

in southern Florida and felt that before the arrival o'

people the uplands were covered by broad-leafed forest.

Robertson (1953,1954), whose viewpoint seems to be

substantiate by more recent estimates of the incidence of

lightning--aused fires, felt that the vegetation pattern was

much as seen today. He also argued that the presence of

pineland an!aeics implies a long period of existence of this

vegetation type. Both agree, however, in suggesting that

the arrival of Anerindians about 2000 yr ago (Tebeau 19A3)

brought about a marked increase in fire frequency an' that







Figure 3. Monthly distribution of lightning-caused
fires in Everglades National Park and prescribed fires
conducted in Long Pine Key pinelands within the park (after
Taylor 1981).


most of thesa fires probably occurred early in the dry

season. The arrival of European settlers on the scene

probably r3sulti in an even higher fire frequency

(Robertson 1953,1954). Besides the obvious effect as an

ignition souc-c (or fire suppression agent), modern man has

other more subtle effects on the fire pattern. The lowering

of water levels by drainage starting in the 1920's has

increased tha time that many vegetation types are burnable

and therefore should increase fire frequency. Lowered water

tables also increase the incidence of severe droughts,

resulting in more fires ia organic soils. On th- other

hand, roais, canals, ani other cultural features form

firebreaks that impede the natural spread of fires.

Present xtent and Condition

The Iiami Rock Rilge pinslands originally covered th!

rock ridge from Iiami to near Mahoqany Hammock in rverglae0

National Park. Davis (1943, see ?ig. 2) estimated the

original area cvaered by pine forests to be about 72,900 ba,

although he stated that this was certainly an overestimate.

The area of the Rockdale soil series can be considered an

independent estimate of the original extent of the rock

ridge pinelands. This has been given as 66,700 ha (Soil

Conservation Service 1958) or 62,800 ha (Leightvy t al.


Today the rock ridge pinelands are almost restricted to

the confines of Everglades National Park. A 1975 survey of

pinalands ani hammocks in Dade County found only 2984 ha or

pinelands outside the national park (Shaw 197r). By 1079

the area had dwindled to 2429 ha (Metro-Dade 1979). Yost of

the pinelands have been destroyed for agriculture or urban

development, Oaly a few areas of pine forest outside of

Everglades iatioaal Park are likely to be preserve! for th"

foreseeable future. The Dade County Parks and Recreation

Department manages three properties with considerable arzas

of pinelani: Larry and Penny Thompson Park (100 ha), ",av

Wells Pineland Preserve (100 ha), and the recently acquired

Tamiami Pineland Preserve (25 ha) which has sandy soil

overlying the limestone (2. Washington, Pade Co. Parks and

Oec. Dept., pers. comm.). There are many other smaller

parcels of rock ridge pinelands in the southern part of the

ridge, but almost without exception they are heavily invaded

by Schinus terebinthifolius, a weedy exotic tree, and are

not properly maintained by prescribed burning (Loope and

Dunevitz 1981).

Although the pineland fire management unit of Everglares

National Park is about 8000 ha (Everglades National Park

1979), plinimetry of a vegetation map of Long Pine Kev

(Olmsted et al. 1983) and topographic sheets of adjoining

areas show tiat only about ~650 ha are in fact pinelands (T.

:aprio, S. Fla. Research Center, pers, comm.). vven within


the park about 500 ha were lost to agriculture before the

land was purchased by the federal government (?art of th"

area known as the Hole-in-the-Donnt).

Today, therafare, less than 10t of the original rockland

pine forest is extant and under some form of management,

There is a very strong possibility that many of the plant

taxa endemic to the liami Eock Ridge pinelands will be lost

inasmuch as only 8 of 17 are presently found in 7verglades

National Park (Table 1).

Eversladesl national Park

The Dinelinds west of Taylor Slouch in "veralales

National Part, known as Long Pine Key (?ig. 2), are thF only

major area of rock ridge pinelands remaining today. There.

is also a suall amount of pinelann east of Taylor Slough

near the main oark entrance. The Loni Pine Key Din.elann

are dissactel by at least six major transverse prairies adi

contain more than 100 tropical hammocks. The pineland

vegetation is described by Robertson (1953), Loope mt al.

(1979), and 31msted et al. (1993).

Evergliles National Park was dedicated in December 19U7,

about the time that the logging of the pine forest begun in

the mid-1933's was finished. The pine stands fond today

are almost entirely second growth, representing the progeny

of Zcll trees. The initial park fire management police for

the pinelands (and the rest of the park as well) was active

fire suppression, in keeping with National Park Service

policy. This initially resulted in qood regeneration of

pine but probably also allowed establishment of larne

numbers of iardwoods and an increase in size of those

already present.

k park service study of fire in the park carried out yv

Robertson (1953) concluded that fire was needed to maintain

pinalands ani prevent succession to hammock. This led, in

1958, to Everglades National Park becoming the first park

service unit to use prescribed fire. Roads wer= construct-e

to divide most of Long Pine Key into ten management blocks,

Details of fire management are described by Klukas (1971),

Bancroft (1976,1979), and raylor (1981) and in the currentt

fire management plan (Everglades National Par'- 1979). After

the initial buris of the management blocks in the lat-

1950's it was often 10 yr before the next burn; since about

1970 the burning rotation has been shortened to about Fverv

5 yr (Taylor 1981). Until 1980 the prescribed burning of

oinalani was :arried out almost exclusively in the cooler

dry season months (Fig. 3). Since then many burns have be'n

carried out luring the lightning fire season.

Prescribe burning is used by the park service to reduce

fuel loads to "natural" levels and as a substitute for

"natural" fire where oresent-day conditions do not permit

the normal pattern of burning (Everglades National Park

1979). A liEficulty in carrying out this type of management


is that ther2 is no explicit sttaement of what constitutes

the natural stat,. Fires caused by indigenous people might

be considered as natural as lightning-caused fires.

It is quite possible that the hardwood understory in the

pinelands af !verglades National Park is today more

conspicuous than it was a hundred years ago. Logging and

the subsequent period of fire suppression may have changed

the balance between herbaceous and woody species. The

prescribed burning program has not significantly reduce the

amount of hiriwools from levels present when the program

began (Taylor and Herndon 1981).


RFperimental Design and Site Selection

Paired plots were set up at two sites

representative of the rock ridge pinelands of Long Pine Kev,

7verglades National Park. One randomly selected member of

each pair wa3 burned during th wet season of 1990 an~ the

other during the dry season of 1980-81. The burns ver-

intended to be is similar as possible except for air

temperature ind fuel moisture conditions, which vary

seasonally. Aboveqround understory biomass anr litter were

simple before burning, immediately postburn, anl at 2, 7,

and 12 mn after burning. The aboveground biomass ind

nutrient stocks of the regrowth vegetation were taken to be

measures of ecosystem recovery.

The sites wers chosen to exemplify both the higher, more

frequently barnel pinelands in the eastern end of Long Pine

Key, and the lower, less frequently burned pinelands to the

west. The choize was restricted to areas that had been

unburned fir a long enough period that sufficient fuels for

complete burns were present. Pineland managsennt blocks T

(site 1) anl E (site 2) were chosen as appropriate sites

(Fig. 4).



0w 0

0 0 0

-0 *
0 4-1
P4 Z

I-I (o

<4 0
4 m0

> 0)rd
I S. *


j r. 0 M-
-r Co

SEl =M 0-

z Ow to)

mm OO

o n

4 '04
4 a0
-rl 0 01

-- .4a

3 4-'


Site 1 contains some of the highest ground in Longq ine

Key and hal been burned more frequently than any other

management block (burns in 8/57, 3/63, 2/63, 12/99, 3/71,

and 11/77). The shrub layer at this site was generally

below 1.5 i, with few hardwood stems greater than 2 cm

basal diameter, and there was a well-developed herb laver.

Site 2 was the least frequently burned management block

(burned 1/59, 4/69, and 1/75) and is in an area where th?

water table occasionally reaches the ground surface. The

hardwoods at this site were larger, often 2-3 a tall, ani

basal diameters of 3 cm were common. Solution holes are

common in site 2 and there was a less well-develooed herb

layer than it site 1. The pine canopy at site 1 was made

up of larger, more widely spaced trees than those at site 2.

The criteria for plot locations were that they ba arhas

of at least 3.5 ha with relatively homog0neons overstorie-

and understocihs. The plots (Fiq. L) were located near

roads (but > 5 m away) for ease of access and to simplifv

the development of firebreaks.

The Burns

The burns were conducted at last 3 a after a rain when

conditions met the criteria of the park fire management

plan, including wind direction and air quality standards.

Prescriptions were kept broad to insure that the burns coul


be done close to the intended date. Fires were set on the

leeward siles of the plots because backing fires arp more

easily controlled and are less likely to cause crown scorch

in overstory pines than head or flanking fires. All burns

were conducted nier midday by the park resource management


Preburn faals ind the material remaining after the 'urns

were collected as part of the biomass and litter sampling.

Wine-fuel-m3istrwe samples (material < 7 mm diameter in 0.25

m2 quadrats) were collected during the burns and dried at

7003. Relative Eire temperatures were measure with plates

placed at ground level, 0.5 m, and 1 m on 12 poles at pos--

burn sampling locations. The plates were made by spotting

temearature-sensitive paints (Tempilaq, Tempil Div., "ig

Three Ind.) on steel plates (75 X 130 X 2.5 mm, 250 7).

The paints had maltinq points of 52, 70, 93, 1'1, 121, 1 ,

149, 177, 203, 232, 263, 298, 302, 316, 343, 371, 3o1 427,

and 45 oC according to the manufacturer's soncifications

(but see 9obbs et al. 1984). The large mass of the Dlates

tends to depcess the maximum temperature registered by the

melting paints so that they more closely reflect

temperatures experienced by the heavier vegetation rather

than maximum flame temperatures. The resource management

staff measuraS ambient air temperature, relative humi'it7,

wind direction and speed, and rate of spread. "irelin-

intensity was calculated by assuming a heat yield of 14,000


kJ/kg of fuel (Wide 1983) and using the following equation

(Brown an? Divis 1973, Wade 1983):

I = Hwr

where I is fire intensity in kW/m, H is heat yield in kJ/kq,

w is mass of fuel consumed in kg/mn, and r is rate of spread

of fire front in m/s.

Atbs2erounl Mass Samoling

Vegetation anl Litter

Abovegrouid understory vegetation and litter were

!sstructively sample! for rry mass. The sampling was don=

in 60 X 80 m plots (0.u8 ha) subdivide! into 12 20 7 20 a

subplots. Within each subplot, potential sampling

quadrats were arranged in a grid pattern. At sito 1 there

were 16 2 X 2 m guadrats and at site 2, nine 3 Y 3 n

quadrats available per subplot. 4t each sampling period two

randomly chosen quadrats per subplot were sampled at site 1

and one quicrat per subplot at site 2. Therefore,

approximately equal areas were sampled in both sites, but

larger qualrits in site 2 plots were used to reduce between-

guadrat variance. Buffer strips 3 m wide between quadrats

permitted movement through the plots without disturbing

sampling sites.

l11 plants ware clipped at ground level exceDt palms

which were clipped at the base of the petioles. In LonT


Pine Key Sibil aLimett, anid occothrinax arqentata usually

have little stem projecting ab-ve ground, and although the

horizontal, creeping rhizome of Serenoa reo-ns is often -on

top of the rack substrate, it is generally unaffected hb

fire and is Eunctionally much like a belowgronnd structure.

Shrubs and palms were collected from the entire quadrat.

eHrbs and litter were harvested only from 1 7 1 m quadrats

nested in tha nW corner of the larger shrub quadrats. Tn

the site 2 plots, herbs and litter were sampled from an

iiditional 1 X 1 a 7uadrat in the S7 corner of the shrub

gaadrat beginning with the 2-mo postburn sampling. Pinp

seedlings were counted in the herb and litter quadrats.

Litter was defined as any dead plant material

iientifiable as to origin, essentially the L and 7 forest

floor layers of the older forest soils literature F(ritchett

1979). generally there is little humus material (U lav-r)

present in Long Pine Key pinelands because of frequent

fires, but in site 2 especially there were occasional

pockets of well-3ecomposed plant matter. A prominent part

of the litter after a fire in these pinelands is the

standing deal hardwood stems: these were sampled from the

entire shrub aualrat. Dead palm fronds were included with

forest floor litter in the 1 X 1 m quadrats as were any pipe

needles draped in the understory vegetation.

The harvested material was sorted into various

compartments, iried at 700C to constant mass, and weighed to


0.1 g. Shribs were sorted by species; leaves (plus rarely

occurring reocodnc:tiv parts) were separated front stems For

the preburn, postburn, and 12-mo sampling periods. PatioleT

ware not separated from blades of palm leaves. Herbs were

treated as a single compartment although record was keot of

all species observed in each plot. Litter was separated

into pine aia non-pine components. In the prbnurn sampling

only subsaomles of litter were separated and the

proportions were applied to total dry mass. At later

sampling periods pine litter was further subdivided into

needles and other pine and the non-pine litter was

categorized as herb litter, forest floor shrub litter, inl

standing dead.

Por nutrient analysis, vegetation and litter from three

adjacent subplots were bulked by tvoe and a large subsamle

kept. In general the categories were the same as those

reacrded for mass, except that only two or three of the most

important shrub species in a given set of subplots were kept

separate and the rest were combined.

The diameters at 1.5 m (dbh) and height of all pine trees

in each plot were recorded along with the number of standing

dead trees (snags) and stumps. Regression equations from

the literature were used to estimate biomass of the


Postburn Ash

Accounting for the mass and nutrient content of the ash

following the fires posed a special methodological problem.

The ash was :ollaetel by two methods. The first consist-e

of placing four Petri dish bottoms (9 cm dia., 21 mm dlpth)

under the litter before the burn at 12 nostburn samnlina

locations (oae in each subplot). Immediately after the

fire, covers were put on the dishes and they were taken

back to the lib. The contents of the four dishes from each

location wera caobined to form a single sample. The

aivantag2 of this method is that the samples can he

collected as soon as the fire front passes; a sudden rain

does not prluale sampling. The disadvantages are the

relatively small area sampled and the difficulty in nlacina

the dishes uader all the fuel, especially in an area wit- a

substrate as rough as Miami oolite.

The se;Bon method was to oick up the ash with a small

vacuum (Car-Yac, Black and Decker) powered by a 12 V

battery. At the same 12 postburn sampling locations a I.?

m2 area was vacuumed and the material was transferred to

plastic bigs. This method cannot be used once the ash is

wetted. It is very effective at picking up all the ash, but

there is potential for contamination with soil. 'oth tvpes

of ash samples were analyzed individually for nutrient

content, although N was measured only in the vac-um samples.


The contribution of the pine overstorv to 1indrstor7

littermass and nutrients was measured by collecting

litterfall for the 12 mo post-fire period. Dine needles,

bark, and male cones were collected in litter trays

(galvanized greenhouse flats, 0.187 a2, with drainage holes)

placed two per subplot in each of the treatments. Trays

were also put out in an unburned area of site 2 to see if

burning increased needlefall. The trays were emptied at 2-5

wk intervals inl material from three adjacent subplots was

pooled to give four samples per plot. The litter was ovpn

dried (700:), sorted into needles and other fine pine

material, aid weighed to 3.1 g. All other material was

discarded. The input of larger material which is

inadequately sampled by litter trays (branches and s'ee

cones) was measured separately at both sites. All newly

fallen (un=hicrri) material was picked up from a strip ?.5 r

wide outside the dry season burn plot borders twice during

the year after burning. Collecting was done in 20 m

lengths, resulting in 14 50 m2 sampling areas. Brarcher

were separated iato 1-cm diameter classes before oven drvin7

and weighing. I assumed that branch and sepd cone

litterfall was tie same at both plots in a given site.

Needlefall was bulked by plot for a drv season perio.

(Dec.-Jan.) for nutrient analysis because these samples were

least subject to leaching losses while in the litter trays.


The total annual fine litterfall (mostly bark) was bulked by

plot for analysis. The nutrient content of branches and

seed cones was measured on material that fell during a 77 ,1

period lurinj the dry season at the site 2 dry-season plot.

materials from the (three or four) sampling areas on each

side of the plot were combined and subsampled to givP four

samples of each type. The nutrient concentrations of the

2-3 cm branch class were applied to the small amount of

material larger than 3 cm diameter.


Two types of soil samples were collected to characterize

the surface soil (0-7.5 cm) and to detect changes in soil

properties prassat 12 mo after burning. The first type was

samples take from the preburn biomass quadrats in all tho

plots and from the 12-mo quadrats in two of the plots. "en

scoops were taken with a trowel in each subplot and then

bulked to form 12 samples per plot. The sampling spots wero

soaked as evenly as possible in each guadrat, but

considerable searching in cracks and crevices was often

required before a scoopful of soil was found.

At site 1 areas of refdish-brown mineral soil were

encountered La the preburn soil sampling: these araas were

readily apparent after the burns. The extent of these

pockets of "Redlind" soil in the site 1 plots was estina+ed

by line-intercept along the 20 m borders of all subolots


(n=31 in easi plot). Because the preburn soil samples wero

heterogeneous with respect to the type of soil sampled, a

second sampling restricted to the Redland soil was done.

oomoosite samples of 10 scoops of mineral soil froa

throughout each of the subplots in the site 1 Iry season

plot were taten 12 m3 after the burn. In the adjacent area

which harnel 5 yr previously composite samples were

collected in a similar manner along 10 m segments of a 120 m

transect. 411 the soils were dried (700C) and sieve~ (2 mn

screen) before analysis.

Tissue and Soil Analvsis

Subsamples (0.5-1 L) of the bulked vegetation and little

material for chemical analysis were ground in a largP Wilev

mill, thoroughly mixed, and store- in polvathvleno bottles.

The ground samples were retried at 700C before wigqhinq.

Nations (Ca, "g, and K) and P were measured on 1.i0 g

samples placed in 50 ml Pyrex beakers and ashed at 5000C for

at least 3 hr. The resiane was dissolved and then brought

to dryness first with 5N HC1 and then concentrated !Cl

before filtering (Whatman No. 42) and bringing to 50 ml

final volume with 0.1 R C1, Calcium and ng concentrations

were determined by atomic absorption, K by flame emission,

and P by automated colorimetry (Mitchell and Sh',? 1979).

Nitrogen was analyzed by a micro-Kjeldahl procedure usia7

0.500 g of litter and stem tissues and 0.250 g of all other


materials. lamonium was determined by automated colorimetry

(Anonymous 1977), The cation and P content of the postbhrn

ash was determined by similar methods except that the

residue aftec fishing was dissolved in larger (107-20X)

volumes of acid because of much higher nutrient

concentrations. Following Kjeldahl digestion of the ash,

ammonium was measured by distillation and titration.

As a chaek on analytical procedures, replicates of a

single leaf tissue sample were run with each set of dry-

ashed s3mples. The ropliration was good, with co-ffici n's

of variation of 4.0 3.57, 3.0O, and 2.u' for K, Ca, "7,

and ?, respectively. recovery of K, Ca, and P from four

samples of Nitional Bureau of Standards pine needle tissue

averaged 81X, 1181, and 98% of the values listed for these

elements. There is no published value for 'g. The 7.,.S.

pine needle tissue was run with each set of Kj1ldahl

analyses and always was within 35 of the nominal value,

Soil pH was measured in a slurry of 25 ml of soil an! 50

ml of water eguiLibrated for 30 min. Organic matter was

measured by the Walkley-Blick vet oxidation method on 1.0 q

samples, Citions and P were extracted by the double-acid

(0.05N HC1 3.025N 12504) method and analyzed as for plantt

tissues. Details of methods are found in Mitchell ana Phua



initial _Veetation Structure

Pins elliottii var. densa was the sole canopy tree

species in all the plots. Both sites had apparently been

logged not long before the establishment of the oark;

numerous stumps still remaining have resisted iore than 11

yr of exposuca and several fires. The density of pinp traes

at site 1 was about half that at site 2, but the mean hasal

area per tree was about twice as large, so total basal area

at the two sites was similar (Table 2). The trees at si'- 1

were also tiller than those at site 2. "any o' the larger

pines at site 2 had misshapen crowns and probably represent

cull trees.

There weca very few pine saplings (< 5 cm dbh) at site 2

and none at site 1 (Figs. 5-7). There were also very few

pines <1.5 m tall that had grown out of the "grass" stage at

either site. Before the burns site 1 had 1.2-1.5 seedlings

(no stem Plongation) per m2 and site 2 had 0.1-0.3 fTahbl


k total of 12q taxa wers observed in the plots during the

course of the study (Table 3, Appendix A). This is a slight

underestimate of the number of species present because some


SE m CN co in
4- a a co o
0 -4O H


0 0 i-
4Ja 1- vm
*0 --a r in in I
0 Z D m C\ m cN
-4J+ 4J
IN M -

oC 0 in I N 0
4J- .0 .
4-z il E!a H m r ;
0-4 a '- IN IN -i -I

(a0 H M * C C
(4 40 (N CN m n

4- ) 0 o 0-
0 0 m a
4M Gr11ia aT in IT in
U) M # + Cm LO i-T 1

0 0 0
0lu o o

4j (a 0 a'a'a.o 0 m
C 4-1

Li-S -
la) a m ) m ) o m r-
'a e 4-1 -H a


4 *0

U )
n 0 M o n m
H o

r II(1
4-1 .-I
(B *d

LL 1



s 10-

0 5 10 15 20 25

-0 -15 20 25
I t.

O -

0 5 10 15 20 25

Figure 5. Size-class distribution
the site 1 plots.

30 35

of pine trees in


60 -



45 SITE 2

40- 11

I 35


L 25

0 -


10 1
5 -

0 5 10 15 20 25 30

Figure 6. Size-class distribution of pine trees in
the site 2 wet season burn plot. Unshaded bars represent
trees dead at 1 yr postburn.


0 5 10 15 20 25

Figure 7. Size-class distribution of
the site 2 dry season burn plot. Unshaded
trees dead 1 yr postburn.

pine trees in
bars represent















O *




'a o



C.) H
4 -1 ) 0

U) H
O '


4-) -,1


(1 a
3> a--
0 H1

asH a41
> 3

C 4-
0 0





(I 1H


m n




uau Umanne co
r-I m r- -p in (N

HIfl 0(N 0
^ ^pr^ f rim ~ o C


Ea 'o >1
.H 0
0 Ea (a In 0 aa 0

ED U 4-0 4- 41 4-
a) 0 -I 3 > r r a

uncommon spazies were probably missed and because a few

herbs were identified only to genus. Within the herbs the

Asteraceae, Poaceae, Euphorbiaceae, and Fa'acea- were

particularly well represented. ?ive shrub species arT

members of the Babiaceae, which contributes an additional

four herb species. Other important shrub families

reflecting tie tropical origins of the flora include the

Arecaceae (piims, 3 spp.), Anacardiaceae (4 sop., including

one exotic), Sapotaceae (3 spp.), and Myrtaceae (3 sup.).

The overall soecits richness dii not vary much hetve-n

sites: however the site 1 plots had higher herb speci-s

richness and the site 2 plots had higher hardwood species

richness. This pattern may be due to characteristics of

the substrate and the close association of the site ? olnts

with hardwood haamocks that aan serve as seed source's: it

could also be due in Dart to more frequent burning o' site

1. Some herbs in site 2 are restricted to the w=tter

microhabitats in solution holes (e.g. Cladiun jamaicpnse

and Thelzpteris kunthii) and some grasses and sedoes at sit"

1 seem to be found only in the patches of ?edland soil (e.g.

hznchoshEorE a E bulais and Desmodium lineatsm). The

hardwood species found in site 2 and not in site 1 are

mostly species characteristic of lower, wetter areas (e.g.

chersobalanus icaco and Ilex cassino) or species found in

nearby hammocks (e.g. Ccccoloba liversifolia and Ivsiloma



The pine overstory clearly dominates the vegetation in

terms of bionass (Table 4). The understory vegetation, in

which the hardwoods are the dominant component, made up 6f

or less of tie total aboveground biomass. The understory

vegetation is much more important in terms of nutrient

content because so much of the overstory biomass is wood,

which has low nutrient concentrations.

Although herb species richness was due largely to forh

species, grlminoais in general contributed more to hbrb

biomass. Prominent grasses include Schigachvria rhioemat'li

(a South F.icida endemic) and An rop2aon cabanisii. Tho

major exception to this was occasional patches dominated by

the fern PtEriLm aguiliaum at site 1. The relative

dominance of the understory shrub species as expressed by

percent biomiss is shown in Figure 8. Puettaria scg-hr wpa

the lominint species at both sites, but less so a+ sit= '

where there were more species. At site 1 Dodonaaa viscose

was the second most important species, a position held by

Evrica cerifers at site 2. At site I the top two species

account for about 55X of the shrub biomass, and at site 2

only 41g. The palms Sabal nalmetto and Sgrenoa renens are

among the top ten species at both sites, but S. reoens is

more important at site 1 and S. palmetto at site 2.

occzothriqnax ar ntata is rare at site 2 an' is nuch less

important than the other two pals at site 1. The biomass

relations of the palms are a reflection of their densities

r0 lco

* m-1 0

(N -V















0' >.


' 1+ 1
4- -4
+ M

,- 4



0 0

a m

+* (N
0' N


T 0

LA 0
+ o
ar r

0 2
U) 0

m a

a 3

m *

re Mh Ln
+ In 0









0 (

44 4J


5 m
C 0 O

O .,.

0 ca

(a C0

m 0

Na C


c 4
4J 0
o o


o o
o -4



m co

+ a0

0 >0
0 0
w 41
0 0
'O U
C 0
0 0

4 C

E-i ai






0 o.5


0.1 WET


0.0 I I I I I I
I 5 10 15 20 25 30 35


Figure 8. Dominance-diversity curves for preburn
shrubs in the study plots. Circles = hardwoods, squares =
palms, and triangles = Zamia floridana. Site 1 plots are
represented by closed symbols and site 2 plots by open

at the two sites (Table 5). Zania Epuila had the eighth

highest preburn biomass at site 1 but was rare it site 2.

Burn Descriotions

The wet season burns were carried out under conditions of

relatively iigh ambient temperatures and humidities,

although the nfel moisture of the site 1 wet season plot was

as low as the dry season plot (Table 6). The site 2 vet

season fuel was such moister than the dry season fuel.

All four burns resulted in 100? topkill of the unacrstory

vegetation, wita minimal canopy scorch. About 707 of ''=

fuel was consumer in the burns except for the site 2 wet

season burn in which high moisture content reduced

consumption to about 50%. The ash resulting from the burns

formed a thin, discontinuous layer on the cock surface un'il

the first posthucn rain washed it away.

The rate 3f spread of the fire front was less than 1.5

cm/s arcapt for brief head fires during wind shifts in the

site 1 dry season burn. The intensity of the burns was

within the optimum range far prescribed fires (73-260 kW/m,

Wade 1983) recept for those brief periods in the site 1 ric

season burn that resulted in the scorching of about 10 of

the crowns.

The averij3 fire temperatures recisteres by the

temnerature-sensitive paints varied significantly aeong t'=

burns (Table 6). The site 2 dry season burn, where the


0 re 4 o 0 aD ew
S0 M N Ni m

4-4 o Ql r

4- U 0 4-)
O r u -1 CU

o 0 H c 4-1
Ou O o o0
04-) C)4-.) (E 04 O U) ev
0 ft 0 U o uo m
'0'0 ,-i m C ) CD
C'o aC) -i
U) : 4-1 H- ( .

a .0
fo a, (1) (o

m m
>H ->

4-W Q4 W
S0 00
0 U r-I 4)
.(-1 0 0
ol Un 04 a 1)
S-i ( 4 co Lfn )N iLn

t -on
4J1 4-1 (D
0 00
0 *H 4-1

NEm m
0) C0C 04 Ln m a
r- a) an m r
0O En Cv
0 4- 0) Q

on u
ri' 4-
) 0

0 4-10

-4 > (0

ll Q) N 0

0 0

S0 0 04
O.4 0c r P: 0

Q 40 a d)
a)~U a I o -ic


4J 0

O .i 04

14m M 4.
>1 CN H
1 a4

Q) O0

-4- ) 00 0 a)

04 D0-- O r-I

0d 0
0 O *

l rO J 4 -H1

rd-H tW (a
-H .-
SrlC V I

C -1 4+ (a
En -H X 4J L4

0 4 -1 En

H 0 Ci

NC.) 3

04> *0
V U0 0

dB rO *CC7

o 40 -ri -H OH
O1 00 4- 04

0 Z0r-
00 ->iU

O4 rd O
0 1 >) 0

-I -r) c 0 -4-r
3 > 10 C C 0

-ri 4 o En M N
4 0 -H>c

U rl 34JU 04
En 9r- Qa) +; FO

) (0 ) 4 0
a0 00m

0 14
3 0

0 ON
3 f

< m
0 CD

4 N
3 0

P m
N 10

In ~o
~cl vl



3 0


0 0

m (M


0 0

a\~ 0r n
r- L m f

i0Oar CO
mr fiir -

^ I*

o r-


O~D m

m3 m

greatest amount of fuel burned, had the highest

temperatures; however, the ground-level temperature was not

significantly different from the site 1 wet season burn.

The ground-level temperature of the site 1 dry season olct

was lower thin the w-t season plot, probably because of

lower ambient temperatures and the more rapid movement of

the fire over th? plot. Temperatures decreased with height

above the grannd, although this pattern was least pronounced

it the site 1 dry season burn.

Fire -ffects on gUn.erstory ass and Nutrients

Initial )istcibution of Mass and Nutrients

The relative Jistribution of the preburn fuel mass amonq

live understory vegetation, understory litter, and pine

litter was similar in all four plots (Figs. 9 and 10). ?Pi.

litter account-3 for 50-70L of the mass and liv2 vpegtation

for 251 or less. The distribution of N and Ca was similar to

mass, but Mg, P, an3 especially K wero relatively high in

the live vegetation. In fact, 75-80 of the K in the fu1e

was containal in the live vegetation. Potassium is easily

leached and gaickly lost from dead tissues. Within the live

vegetation hardwood shrubs were responsible for > 507 of the

mass and nutrients, more so at site 2 which had not b'rne~

for a longer time.


- Ii


N zZ I
o'n ,

- W
- C


._ -P



o 0
o o
o w

I- -H
z p

L r



S >1

o N

o 0


o i





a- E

s I
4 0
S2 a




o 0

o oa
0 ,
U (D


8 5
- m





4 2

Ash Collectian methods

There were no significant differences between postburn

ash colleztipn methods for mass, nutrient concentrations, or

nutrient staiiing croos in the site 1 burns with a single

exception: liighr 3g concentration by the Petri dish methfo

from the dry season burn (Table 7). There was, however, a

consistent pattern of higher mean nutrient concentrations in

the ash collected by the patri dish method. A light rain

falling near the end of the site 2 wet season burn proclud=d

the collection of ash by the vacuum metho feor that hurn.

In the sita 2 dry season burn the estimates of ash mass ani

all nutrient standing crops were significantly higher for

the vacuum 2eth3]. The tendency for higher nutrient

concentrations in the petri dish ash was also found here,

with only Mg significantly higher. Apparently at site 2

substantial aaounts of unburned humus and/or mineral soil

were pickz1 ap by the vacuum. This would account for bo+h

the higher mass and the tendency toward lower nutrient

concentrations. Because of the larger surface area sampled

by the vacuui method, it produced smaller coofficients of

variation for mass and nutrient concentrations in 14 of 15


The dry mass and nutrient cationss and P) values froi the

patri dish acthoi were used to estimate postburn starling

crops because they were available for all the burns anJ

because vacaui samples gave over-estimates at site ?.


**>o N *T i^ co
r4 14
fC N- 0 o 9C 9 0 0 0 00 N CN N -
C N C ( N *
0 O 0 0 0 0 0 O x a a t s 1
0+ -H -+ i +1 +- +1 N - (N
0 + a N Nn aN N

44 II ^ Ci f r M 1 + | +| +U 01 9 0 0

C .. m ,o
02~~N .9 U, U, 0 N

Q) 4 JI,,,
0 - - -' (N ^ N 0 U N U

a) 1
Cn > I-9

Wc 104 .
U 0 o -N O 0 N0
QJ o Ao *m rs N

H C N N n N N N E I 9 .9
64 C N N U, N UN N N 9U, U, (N
CC + 0 z4 - ( N c f
S. .
0' -, +r +9 +| +9 +9 +9 -

HC (0 N U N -
Hi- C -l P- N N N( U 0

o 0C S o
C," o1 f^ n 3 m in CT\ ^i o 01 O

0 -0
0, N (Ni N N N N 9 CCT C .
02 Cj - (N 0t (Ni -- C 9 +9 + + 9
tfl ^ 0 0 0 N O N .
rlCV cl I n p^ ,- 01 3 n f d o o

S . . + 0 U N
4 '. 9 a 0 C C C IO^

(U U > U U -< n (N + U, (
024 ^J (N N (N (N (N NI (N +4
S0 09 C0C N 0 0 O
4444 N N 0 N N (N NU
02I N N) N N N (N (N D a O

NJ No 0^ '' -' N N
+91 S N 19 0 .9 0 N .9 N .9

020 9 '0 &
02 i 0 I .9 I I CM ++ N +l (N U, +
Qi Z 1 0 0 9D 03 -9 NU N ND U, N
N O l +i N +i l9 I M [- l n m o 9 9 + (

O 0 N U,0

wu w
(9 NN N U U N ON C .
E N U N U IF 9~ 9 NI

'C) ., 42 o 2 2 *I

0 u +9 +9 + +9 +i +0 +9 U, m (N

I+iO N NP N^ No N- (N N- +9 o 0 +9
aO U, N N N N 9+ U,
9+ (N .90 (N
(N Nl 0 U . -o- c
aC >uIC +9 +9 +9 N N N -
(990. Ul U,' NO N NO M (NO.9
02 Q C ) I r CT co ul r~

C0 U'' C U' '
-j02 .9 C) C +9+9) C 0 -5) E e C

+9 +f1 Cr '4 14 C CCC + .9 +9. 39 C C
. -i ,? U' C^ E' C^ N*- U' C" U ^
rl ., I- I 3 D -
a vll X: E. +I U J U +1 *

024 w '9 '9
04 in r a u
fa n3 o

00 C U
C (99 +4 J >. +9 .99

C' .9 3) +4 ii) 04i
~ 3 n 01 3 CJ 3 kj

H O C)I +9 -I -IU +IL

r^ oa a Q

.fl +9 (N (N +

02 .4 E .

Nitrogen in ash was estimated from the mass an? nutrient

con7entratioa data collected by the vacuum netheo. Since

these data iare not available for the site 2 v t season

burn, the meB n J concentra ion for thm site 2 3rv s-ason

burn was applied to the wet season oetri dish estimates of

ash mass. Tie estimate of N in the postbcrn ash at tbh sito

2 dry season burn is undoubtedly an over-estimate and

thicefors qi4?s a conservative estimate of the loss of for

that burn.

Postburn istribhution of !ass anj Nutrients

lost of tie litter (except for the heavier pine branches

and cones) was consumed in all tho burns (iqs,. 9 and 101;

in the site 2 wet season burn considerable amounts of th?

oartiallv lo3ompos lovwr layers of litter in '~oressions

wrea also left. Yost of the preburn vegetation di- nomt qrn

ercept in the s[ta 1 wet season burn where sliqhtlv morn

than half wvs consumed. Herbs were almost completely

consumed in the burns (exIcpt for petioles of teridiun),

but shrub stams ian most leaf material remained. The shrub

stems that remained upright became the standing dead

compartment of the nostburn litter. The blad-s of paln

fronds were often at least partially consumer but the

patiolbs did not burn. rhe 78-95 g/mn of postirrn ash

accounted foc 9-2;- of the mass remaining after the brrs.


The losses of nutrients from the litter and vegetation

were roughly proportional to the losses in mass (pies. 9 anP

10). one-haLf or more of the standing crops of the non-

volatile elements other thin K were found in the ash after

the burns. A large proportion of the K remained in thi

dead, but unconsumed, vegetation (mostly shrubs). The

nostburn distribution of N was similar to the nosthurn

distribution of mass.

The overall losses of organic natter from the ecosystem

ranged from bout 1-1.5 kg/m2 (Table 8). The losses of 7

ranged from 5,7-9.5 /ma2. There were no sijnificane

differences between preburn and postburn standing crons of

the non-volitLle elements except for a loss of K in the sit?

1 wet season burn. This represents one of 16 tests (at 0.0~

level) for nonvolatile nutrients and may bh a falss

rejection (i,e, a tvoe I error). There is li'tl- re:-so '

exoect detectable losses of K in relatively cool fir-s sIuch

as these,

Postfire Pecoverv of ahss and Nutrients


Although abovaground plant parts were killed hv the

burns, I observed almost no plant mortality. r'e of *he

most striking feitur-s of th= postburn period was the rail

reappearance of green tissues, especially herbs anl nalis.

Within days aftar the burns fresh blades of grass apnear"-


c0 -rHl t r H0 f fl-I 4CN C'i0]'0 f l'LAC 1C r-4
-1 0 y-4 P 000 o0- r- CNN O- r-1
0 0 ( 0) *n o M . . .. .
T 54 >4 rWH -Or 000 000 a 'N CC 000
0 0 n +1 +1+1 +1+1+1 +1+1 +1 +1 +1 +1 +1+1+1 +1 +1 +1
Sd Q W0 n n 0CN me -e1 N -A Ln fl
aNiLn mClo :1;o . Lo
0 .*>i m n o mmo * * mn ..
rr ~QH r i 000 HH- NN I CN -0C

0 0r

C!-)V -H
0 020 4

4. C-C CCNO0 000 i-irl(N NMWo O0mt
C 4VC m t0 n . . . . . . . .
r00 4-J 0 +1 +1+1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1
O +1 4J 3: 0)oC r43 m a- 0 0 %D 0cr- CN ) .N
C 0 * r-0 1m- *** NNO
0 0zm a a **0]) 0 o * **
024Jn -1 00 C -4-4 40 C-C C -NC)J

e -1a
0201 0

rO tp n3
02 4

0 c 0 P-4ro ri i N o or r- r- OD~ w r %o00

[ fa0 r o ** C D .. .
4 >4 qnv 000 000 3ooo Coi4 000
0 o 0 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 + +1I +1 +1 +1
1C m M -4 C- mmC0- L* ** No 0mmi r 00o-
If I] r.. 0000 rz--p . CN Or lC O

-H a)
c 000 000 O 000
Cr 4-)--
40 -l a) *
-H o -4r

0C 0 10 --4 rH- w )-4
44 OI 4- 000 r-(m mom m3Crn 0om1-
C- 04 00 0 000 -4o- oNc 00 -i
4 4J 4 UO N m 0 0 0 0. 0 0
2 4J -4l Q 1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1
00Q) 4-4 COOD CfwrD m00zzp oVwLf-i 0Oi 0m r-im 1--
0a r* C- NCJO Mc *o nO

S> z i- 000 -4 r-I0 r- r-I -O 0

0C r-I

Q4 02
14 C -H

4 14CS- CC4 C14 Ce c -i
m Ul -U w1 p;3 01 H :: 3 Pi 3 M H :J U) t 3 01 : 4 3 V2
U 2 na C 0 3Cf 0 0 2C f0n -0-:(a -H30 3 E3 0 2 -Hi l0a
0 2a r02 O 4 0 .2 4- 0 434-C +0 02 f O 30.0 41 0 M 0 2 l0O
02 Q -H 02 4 0 W (n-? a ) m 0 2 w n w00 m -H0(D(a- (a 02 ww
i o o i o 0B 1440 p 0 CU 1o 00 0p0 Zo40
.0 >iO C-0. 4- Hi PI PO p0 4- A. PC r-4 l P4P tnP4 P4
10 1 C -HJ C 0 n0 r0

from the ground and palm fronds pushed out from the stem

oices. f arwoon recovery began somewhat later.

Essentially all the recovery during the first year was

ieu to snrouting from belowground parts. Among the

hardwoods only Dodonaea viscosa and Rhus sooallina hai any

seedling regeneration, but this contributed an insignificant

amount to tta total recovery. The seedlings probably came

from seeds present in the soil or litter before the iurnm;

both genera are known to have fire-stimulated germination

(Ploy 1966, larks 1979). 7o data on reproductive activi'v

were taken, 3ut luring the year following brr.in7 I o'bserv

flowering of every herbaceous species Dresent in t'he plots.

Within a aoith after the site 1 wet season burn inellia

caroliniensis was flowering and by 2 mo 15 aliitional

snesies were in flower. productivee activity of palms was

not noti.Ze-bly different in bu n unbrn ai na? ar as. Mo t

hardwoods dil not flower during the first year after %rning

except for t1e weakly woody Lantana devrossa, 1orinda rovoc,

and Zcoton line ris. Occasional individuals of Psidium

lon2!ai2S 3lrsonima lucida, Iodonaea viscosa, anI Ficus

zitrifolia also flowered within the first year.

Harbs reached their preburn biomass levels within 7 mo

following dry season burns and hy 1 yr after wtt season

burns, even though the 2 mo recovery was grea +r after wV

season burns (Fig. 11, .ppendix B). In all four of the

plots palm biomass by 7 mo was statistically


indistinguisiable from preburn amounts. Hardwoods recovered

more slowly than herbs and did not approach their preburn

levels by thE eni of the first year. Recoverv a' 1 yr

ranged from 19-39' of the orebturn biomass, Certain hardwood

species sprouted sooner than others (D. viscosa and 2.

c2Eallina wece two of the earliest). Hardwood biomass

increased at each of the sampling periods following wqe

season burns, but showed no increase from 7-12 mo after th

dry season barns (Fig. 11). The overall pattern of r;coverv

indicates that koril to August is a more favorable qrowi-r

p~riol than the rest of the year.

Before the fires, stems accounted for most of the

hardwood biomass, whereas 1 yr after burning leaves

accounted for equal or greater amounts than stpas (Appendti-

3). ar.wn33 Leaf biomass recovered to 32-68- of t.h

prehurn biomass but stems reached only 12-2' oi' *he

original amounts. The preburn stems, of course, represent~?

the wood increments of 3.5-6.5 yr of growth.

The relatively rapid recovery of leaf biomass combinoI

with complete recovery of herb and palm biomass means that

the functional capacity for photosynthesis recovers faster

than tha structural characteristic of biomass. If ons

assumes that the sum of herb, palm, and har-wood leaf

biomass is a measure of ph3tosynthetic capacity, then t"- 1

yr recovery of the vegetation for wet and dry season burns

is 82 and 93t at site 1 and 74 and 58' at site 2. The

corresponding recovery of total vegetation biomass

(including hardwood stems) is 55, 63, 41, and 27?.

The above7rouad net primary production of the unerstorv

for the first veir after burning can be estima-ed ty adding

the litter produced during the year to the standing crop op

biomass at the and of the year. There was no palm leaf

litter an! little hardwood leaf litter produced during the

first year, iad this litter was not measured seoaratelv from

the hardwood litter resulting from the burns. There were

substantial amounts of herb litter produced hv I "r

(11.7-39.5 I/a2, Appendix B). Fstinates of not production

are 156 and 166 1*m-2+vr-I for wet and dry season turn plots

at site 1 ani 20) and 144 g*m-2*yr-l for wet and dry season

burn plots at site 2. In addition to the few hardwood

leaves shEt1 y 1 vr and the loss of mass by herb litter

through leaching and decomposition, herbivor' is also not

accounteS fcr vb this estimate. Grasshopnprs anq

lepidoptcran larvae were the most conspicuous invertebrate

herbivores; the caterpillars of the echo north (Sqirarri-

echo) grazea the young leaves of Zamia Eamila vary heavily.

I also witnessed four white-tailed deer feeding on the voung

regrowth of 3a0 3o the plots.

Because tie herb and palm biomass 1 yr after burning wr-'

not significantly different from their preburn values, th=re

is little evidence that the season of burning had an effect

on their recovery. Variability in the data nake it

impossible to detect small differences that miaht have a

cumulative effect after several burns.

Rardwoods, on the other hand, showed arkd differences

in recovery among the four plots. The absolute recovery of

hardwood bianass was greater at site 2, but that site had a

nmuh higher initial hardwood biomass. At site 1 the

hardwood recovery was significantly lower following *he wat

season burn (29 ;/ma or 20) than the dry season burn (f?

g/m2 or 39f). At site 2 the trend was reversed, with

greater recovery after the wet season birn (137 7/m2 or I"1)

thin the dry season burn (78 g/m2 or 18!) Th? most~ li'=lv

explanation for this apparent inconsistency is that fire

temperatures had a greater effect on recovery than season of


The recoverv of nutrients in the undzrstorv veetation

followed a pattern very similar to that of biom~ss ficn.

12-16, Appealix 5). The maia difference was the somewhat'

faster rate of recovery because of elevated nutrient

concentrations in the young regrowth tissues. The recovery

of Ca, however, was delayed somewhat relative to biomass

because the 2 io tissue concentrations of palms and

hardwoods were lower than the preburn concentrations. Herbs

in particular often reached preburn levels of nutrient

standing cross earlier than biomass. For exarple, ?

standing cron in herbs was not significantly diff-rent fro?

the preburn amount for both dry season burns at only 2 mo






2 100




S I 150





0 2

12 0 2

Figure 11. Postburn recovery of understory biomass.
Bar on left shows preburn values.


cU 1.4
0 1.2

WZ 1.0
0 0.8

Z 0.6







o 1.2

0 0.8
Z 0.6



3.3 7



12 0 2

8 12
8 12

Figure 12. Postburn recovery of nitrogen in understory
vegetation. Bar on left shows preburn values.

0 2 7 12 0 2 7


0 2 7


0 2



12 0 2 7 12

J 0.167

12 0 2 8 12

Figure 13. Postburn recovery of phosphorus in under-
story vegetation. Bar on left shows preburn values.


E 1.2-


N .-

O 0.6-





o, 1.0-

2 0.8-




12 0 2

8 12

Figure 14. Postburn recovery of potassium in under-
story vegetation. Bar on left shows preburn values.


[1 - - -

0 2






7 12


12 0 2

8 12

Figure 15. Postburn recovery of calcium in understory
vegetation. Bar on left shows preburn values.



0 2 7




E 0.25

( 0.15
< 0.10 HERBS

0.0 5

0 2 712 0 2 7 12

0.788 SITE 2


0.2 5

0 2 7 12 0 2 8 12


Figure 16. Postburn recovery of magnesium in under-
story vegetation. Bar on left shows preburn values.


postburn; thi standing crap of K in herbs in the site 1 wet

season plot also reached the preburn level by 2 mo. Tn one

case (Ca, site 1 wet season burn) the standing cror was

significantly higher 1 yr aftar burning than before th-

fire. For palms, both N and P standing crops were not

significantly different from preburn at 2 mo for the sit- ?

wet season burn. All other palm nutrient standing crons,

like palm biomass, dil not rcach initial amounts until 7 mo


The pattecn of recovery of nutrient stanlini crops in

hardwoods was the same as for biomass except thit K at 'lh

site 1 dry season plot at 7 mo and the site 2 wet season

plot at 12 no were not different from the prpburn amounts.

The percent recovery of all nutrients in hardwoods at 1 vr

was greater than th? corresponding biomass recovery tecausn

of hiTh nutrient concentrations. The Dercent recovzer7 o

the total -veg;attion nutrient standing crops was greater

than the reco7vry of biomass for the same reason. Th- t

standing crops in the two dry season plots at 1 vr ware not

significantly different from the preburn standing crops,


The standing :rop of litter sass (and nutrients) iurin-

the 1 yr postbura period is a function of the amounts

present immediately after the burns, inputs, and losses

through learning and decomposition. Even though nearly all


of the preborn understory litter was consumed in the burns

(Figs. 9 ind 10, Appendix B), the understorv vSgatation that

was killed bat not consumed by the fire replace, it. In th?

dry season plot in site 2 there was actually more after th=

burn than before, The understory litter compartment

received no inputs for several months until herbacouss

material h-gan to die. By tho 12 mo sampling nerioi some of

the hardwoods had also shed some leaves. At 1 yr some ')

the oldest palm fronds were becoming senescent, buit were not

dead and therefore not 7et part of the unnerstory li'tor.

There was some decrease in understory litter nass diiin

the first two months after burning in all plots (Annen-li

3) From 2-7 mD after burning the dry season brn lots

showed continued loss of mass (because these were wet season

months), while the wet season plots did not (h=cduse it was

dry). "rom 7-12 mo there was an incrpas3 in Pass excent at

the site 2 Iry season plot where decomposition out'acpe

meager litter production by shrubs and herbs. At si*e 1

much of the input to understory litter is attributable to


kt the ead of the first year the standing crops of

understory litter mass were not different from anmonts

present immediately after the burns, except for the si'e ?

dry season plot. however, the standing crops w-r0 h-low

their preburn levels except a the site 2 wet season plot

where there eas a large amount o" unconsumed v=gotation ani

rapid recovery of litter-producing hardwoods.


The standing crops of nutrients in the nnd-rstorv litter

shovel various patterns during the postburn year. Larqe

fecreasss often occurred luring the first posthurn rainy

period, 0-2 no following wet season burns and 2-7 ml after

dry season barns. Potassium in particular showed leachinq

losses of >80-90t of the amount in the postburn litter. Fr

most of the nutrients there was some net increase d rini the

7-12 mo periLd. Phosphorus, however, showed initial losses

and no significant change during the last S no period.

"itrogen and Ca showed the least net loss during th- voer

and were not different from the amounts present immediat=lv

after the burns (except for the site 2 dry season plot which :

lost mass throughout the period).

In contrast to the understory litter compartment, which

had no innuts during the firs several months 1cter birninr,

the pin. litter rcecived continual inputs. ?n.c nertle?

fall throughout the year, but at highest rates during th?

wet season (Pig. 17). Short term peaks in neertlfall ar?

caused by high winds. The annual needlefall was about 320

g/mn for site 1 and 260 g/m2 for sit- 2 (Table 9).

Seedlefall constitutes 75-80 of total pins litterfall.

Pine branches, mostly <3 cm diameter, contribute about a

tenth of tha litterfall mass, and seed cones and

miscellaneous materials make up about the same fraction.

Some vear-to-year variation in litterfall is to he Ypex'ect-

needlefall during the study period was about vveraqg a-

I .0

, L








Figure 17. Pine needlefall in the study plots and an
unburned area of site 2. Closed triangles show time of burn
and open triangles the 1 yr anniversary of the burn. Absence
of bars means that no sampling was done.

I I I I I I I I 'a I I I I

1.5 -j

? 0.5

- 1.5

S 1.0

w 0.5"-
_J :

z -
LJ :
E 0.5:







r-H m -lm rl r0 nl rm
A in Z3 mn r' -I CV IT

.T d C r OtoOD -1 C a) 0 -4 m
iLHC- yH(N (NH I.-1 N m LA 0
H- -1 -4 -4 r-4 r-q I H -4 H -4




m 0


*H O


0 0


+1 a

rn m

U) 0

4J,4 -I

0 '


5 -

--4 0



r-4N N o- 0

SO r 7 Ho! NH

O( NH mr(Ln mm (NmN
O r, D 00 o 0

vnW o (NL .N M mmm oor No
a m i m N c m c

SCN N c mm
- W r r- -i
0 r oN- N
rH -4H H H-

* i
i L
H U1
+1 +1
n n

m 4

* il r-
N *
(N l (VN
+1 +1 +1 +1
0 00 0 qr

a; 1 r

CN Lo all C) Ln Ch m fi ^> msCN M
[r- (r'- in) OO O()a%-4 m oL r-
N 0 0 0 m 0
r-4 -o o

* 00

m 0

T (N o
(N m

WDm 0
* ooo o
H ..

+1 +1 +1 +1
oC) ) o-

r-(N a
(N m

0 O

*+ +


(N m

C $ C r r.
0 0 0 0
Un E Hi a) H U) ) U) 0

r-H *- H *O U (n 4-H H *O- -I U L 4) 4-)
0'0a u 0 a'o r C0 c o ()a 0c0 u 00
0 ) U) E-p 0 ) U) M Z E O Q) UJ O- ww 0 z E-4 (a CE-a
U) w)-4 ( i H 0 U) ) *. a) -. p o o)-1
m (Z Z0U M UZ XE-1 mZxqU a)za

) 0 0
0 o T3 3

S(N (N (N

) a) a) ) a

u 0E m

o a

o- o-



0Q -




other locations in Long Pine Key (A. Herndon, unpublished

data). The number of seed cones was probably higher than

average, at least for site 1, because there was a better

than average seed crop in the autumn of 1939. The

deposition of nutrients through pine litterfall ranged from

about 2 g/m2 of Ca and 1 g/mr of N to 0.03-0.05 q/m2 of ?

(Table 9). Ine notable aspect of the nutrient inrnts is tF'

high concentration of K and the low concentration of Ca in

pin czones relative to other litter tvpes.

Pine liter mikes up the maior portion of the f'il mags.

The material remaining after a Fir? is mostly heavier pieces

such as branches and cones, unless the fuel is moist (;,4.

site 2 wet season burn). Pine litter mass showed a more or

less continuous increase ercent for the 9-2 mo oerio,

following tih wet season burns where moist con1;tios anI

the nutrizent release in the ash may have 3timula*

lecomposition. By 1 yr all plots were still- far below

preburn levels: site 1 about 50* and site 2 33-40' -o


The nutrient standing crops in pine litter showed' a

pattern very similar to mass with some decrease 0-2 mo

following wet season burns and increases otherwise. The

percent recovery at 1 vr was similar to mass for all

nutrients except N, which only reached about 337 at site 1

3nd 20% at site 2.

The recovery of total litter mass followed a pattern

dominated by the pine litter except for the ldcline between

7 and 12 no it t1s site 2 dry season burn plot causs by! a

drop in understory litter. At one year litter mass was

42-62% of its praburn level. Phosphorus and Ca recovered in

proportion t3 dry mass. Nitrogen was much lower 1han dry

mass (26-37A) because C:N ratios decrease with

decomposition. Potassium recovery was high relative to

mass, esenaiilly at site 2, because of the high proportion

of relatively unweathered litter materials.

TPMact of wires on Pines

There is some indication of a slight increase in the rate

of needlafall immediately following the burns (Oiq. 17),

even though there was little scorching of neseles. The

stress of hijh taeperatnres may lead to senescence of nrePl0

fascicles earlier than normal.

In both site 1 plots there was no mortality of pip? trees

during the year after burning (Fiq. 5). In contrast, a'

site 2, where the trees were smaller, some trees died (vigs.

6 and 7). In the wet season plot at site 2 six trees were

dead 1 yr ifter the burn and in the dry season plot 98 +*res

were dead, including virtually all trees < 7 cm dbh., one

of these treas showed substantial scorching ind n 1al npe"r

to be healthy for the first few months after the fires.

There was evidence of bark beetle activity in all the 9e


trees, but there is no way of knowing whether the insects

were responsible for the deaths of the trees or whether they

only attack] trees that were already dying. T' is cl-ar,

however, that high temperatures around the stcms of Sout'h

Florida slash pine trees can lead to their deaths. Three

years after the dry season burn at site 2 most of the pines

had cracks erudinj resin in the bark of the basal 1-? m of

the trunk.

Under the burning conditions of the eTperimental burns

all Dine seedlings were killed by the fires (Table 105). ms

season of burning has a pronounced effect on +-h

establishment of new seedlings. Seedfall is from S-teimber

into Novembrc; therefore burns that occur in the wet season

before seeffall create excellent conditions for see?

gernination inl seedling establishment. Burns that occrr in

the dry season (after seedfall) destroy the caorrc* ve-r's

seed crop ani by the n-xt period of seedfall conditions ar=

less favorable for seedling establishment. The 198n se53

crop in Long Pine Key was relatively good and the wet season

birn plots hid much higher seedling densities 1 yr later

than dii plots burned in the dry season (Table 10).



o r-i


a i i

+ I

a) 11

S' )




x0 4
tU 1C! -0
D 0 -H

d d

0 w 0)

4'- 04
a o

*a 3

4- 4


0 0
0r ,~

E- i

4) >4 4-) >4
Ct M (
g Q S Q

o 0 0





Soil analyses show some striking differences in th-

properties Eo the substrate between the two sites (Tabl?

11). Th'se differences are largely due to +he oresence of

pockets of tie reddish-brown "Redlani" soil at site 1. The

wet season CLot had 8.21 (S.E.=1.62) of the surface covered

by mineral sails and the Iry season plot B.uT S.7.=1.42).

A very few packets of the Pedland soil were present in the

site 2 dry season plot and none were seen in the site 2 we1

season olot,

The organic matter content of the soils was much h.'br

at site 2 (iaout 46-) than site 1 (about 183), although bot-

sites are relatively high for upland soils. The higher

organic content nay account for the slightly lower nrnhurn

op at site 2. ?reburn extractable K an( M'I to not show ich

difference between sites, although K ray he so3ew'- hi T'-r

at site 2. Phosphorus is much higher at site 2 than sit? 1,

suggesting a relationship with organic matter.

Burn eff-;ts on soil properties remaining 1 yr postbnrn

include an increase in pR. This was most pronounce! in *he

site 2 samples ani in the paired Redland soil samples taken

in site 1, were initial p" was lower. Soil organic master

was unaffected by burning. There is an indication that

extractable P was reduced 1 yr after burning. The amounts

extractable by weak acid in the Pedland soil samples were

below detection limits both in burned and unburned soil.


9 in in o .
,H-i 0
m I in --4 r- -4 L)
4- +1 +I +I +1 +1 +1 +1 +1
O 0 H m 0 -i
,- D -0 I 0 0)
Q L Hl H H N C4N H
NC C' 00 C 0N 0 rH to O
4 a m m r-
4-) en m to 0 N 0
.> n) Q .34 (N H HA H H 3cH
0 M +1 +1 + +1 +1 +1 4 +1
O m r- r' N o m m
,.C ) 5- .. ..
t C H LA O A O O to L A
3HO -0 en L m a n N en
O" L

,'- 00 L.r) o 'o r-.
S-i I I I n a in n
0 0 ..

*H (N N rH
o En .0

a) a) a

w CN r, m0 H
40 W- U
Qo-,-i Wr 44 o - m U O o "
"-I +i +i +i +i +i +i +1 +1 +i
HI'n CJ4 ( N ND O in o N -t
0 )O .
,- .,- H ) H 0 en en en o 0 r-,
H ) .4,+ + +5 +3 +O + L + +a

r--I co co 0 1 O O ) O-
O -3't O o4 o ( ,o 3d L '.0 on o ,
O .
n D -I + +i +i +i +i + +i + +i
C M) 00

H 43 4 N ON N A N ON O

5C)3d H
-1 ) 4(
(a ON In N ON Loo LA
mo p r* :: 0

U 1 4
H41 |l4J m' C C C

rn : cc
WU)UH 0 C 4g ) C >1U C) C) >cU) eH >1[n
(a0 m o4 p4 H 04 04 oioR w Hp4

U) 1
4-1IC) t4)

E5-4 0 0 C
3 ec4 4c Ct>
U'o 00 4-i >j >1 4.1 >1 >1 .3 >,
CG4J-m1 H C) 1 1 C) C- 14 C 3.

H C) ]

Extractable Ig May be somewhat higher after burning, at

least at sit? 1. Burning increased extractable K in both

types of soil simples at site 1 but the increase was

statistically significant only for the Rodland soil. At

sits 2 th=re was a .dcrease in K after hurning. This may

result from aore rapid uptake by the vegetation at site 2.


Pire-caused Losses of Nutrients

One of the most obvious effects of fire is the removal of

organic matter. Nutrients are also lost by volatilization

or articulate emissions. Numerous studies have at+emnote

to ai-sure tiass losses in several kinds of ecosystems. 'n

prescribed b.ining situations it is possible to sasnl"

standing croos before and after the fire, as was done in

this study. Wit1 wildfires the approach has usually been to

compare starling crops in the burned area vi-h nearby

unburned analoqs (e.g. 3rier 19751. Spatial hferogenoi'

of fuels inl the collecti~q of nostburn ash co3se Frble's

for field studies (Prison 1980). "or example, in a lic

prescribed birn in southeastern coastal plain oine forest,

Binstock (1979) and Nguyen (19781 measured higher Tean

values of mass and nutrients after the burn than before. In

a study of fire in heather ecosystems in nqglan! Svans and

Allon (1971) gave up on field measurements and performoi

artificial barns in the laboratory. This reoresents th~

other common approach to measuring losses during fires.

Simulated burning has been done both as open ignition or in

a muffle furnace. The additional accuracy in measuring'

losses, however, is at the expense of the realism of the

burning conditions.

Table 12 presents the average mass halanc? for th foie

burns in this study along with data crom several other

southeastern pinelands and a few other ecosystems. Th-

table does not include examples of slash burning (e.g.

Harwood and Jackson 1975, Evel Pt al. 1991). Comparisons

must be tBmnerei with a degree of caution because

methodologies anI conventions differ somewhat amoag thb

studies. Ti sone cases the fuel rPnrZsents only the fore: t

floor (litt=r) and in others litter 'rd veq-mttion. T '

studies h7 3ough (1991), Kadama and Van Lear (1o ?inh'=r

et al. (1932), Debano and Conrad (1978), and grier (1971

involved field simpling of ash, Debano and Conra vacu e--'

the litter and ash in their stuiv of chaonrral. 'o'lu-

picked uo the ash "by hand" and stated that som? ash was not

collected. The study by rough (1981) '1als with the

situation most similar to the "iami Fock Pidqg oinelanos.

The overstory was mixed stands of slash and lonqleif oines

with relatiTvly Iense understories dominated by Sersnoa

rsgons (saw nalomtto) and lex glabral (gallherryv.

The fuel consumption in the Long Pinea ey burns was

within the ringe of the other pineland burns [Table 121,

The lower consuaptiDn (both absolute and percanta e) i ;

three of tae other southeastern oineland burns is idue o

hiqh moisture content of the lower layers of the for-qt


m r
O 0
I-I -4
0 41




11) "




r4 ()


, r-i

c o


Cr J (U


C 4- 01
o) W -


o N NNo N o N N N N
- N 1 N

o N
0.4 i.3


0^ C
C: B




U) 0


0 0

- 0 0 -

O 0 4
N 0
0 N 4.)
U .3

4 4C
-o o w w o c -0

o0 0 m w o
0Cl-l N N T -N N

OWN N No -- Na .4 -IN
0)' 0- 0< *4-'^ ON 0) OOV c 41)
c o~ 3O O -r c' L
00)0 0 LI1)wI0 0)0)0) 00)

0 a. o -
N0)1) S- B 1Uat^-N [dN-.- 44Q..
dLI-.0N''.4 440)NN E 0) 1l~ )4" ONN c
q W W V V^ -4 Q 04

* -I 3ONN44O.ic c.4 a 4
0 4 N 4- N 0N .0 044
4 3N

+ I- + 0) +
-~ -~ 3 .4 .4 .4

1 0
T \

I 1 0 1

^ C

m< k

tf r~ ^-

I I I 0



1 =
0 04

SS "
N N^

44 ON
0 440


.0 E-




4 -, ,-



0 E


C: 3




m 0
11) 11


o S
( 0

m Ia

-4 .1 145

m-C 0 m C -3
m 4 0 14

10 4, 0 C M 4
-o .4 . (0 W C IN.

441.1 .CINC.. INC. *.40.C

40 0 04 CC C14
C. 44- 14 r *
iq ffl ^ C a\ V-

CCrCNC4 j C 144,N4.4V4.4NUC4,N
443 34 1 n C I 4 6 O
44 4,14-4, V^ Vt 0-04-
141>1>.4.41414401414144444, 4

4- --4 C 44 n 0 >4 14 44 4 4, l 4 4 4,
1 0 4, u0 -4 C i 14 4114> .-4 o44 1444
-^ -it 4-n IN 4, U 4, .4 c) c4 J o C. n
.0- fl~i4 j l fl C, if O 44'- 4444 Cl

*a'-xi--^- E- H)^(O-TIC-
>.4S -4 >5 3 IN3 IN U U

o o o e

r- n n o i

0 0 N IN

c o ao r w c
(nm V (N *0-

0 'f < *^ CO N

0 00 0 00




C 44

Pi Lo

aa o0


0 II3
C fl K

3 4 -4

,, ,,,,


P N .o'



0 0 0


P 3



00 iP


N~ 01



floor (Sweenav and Biswell 19F1). Wildfires under oxtrerm

conditions 3in oF course Lead to very large losses (Grior

1975) rasslani fuels (Lloyd 1971, ChristPns-n 1977) also

have high percentage consumption rats.

The losses of 5 seem, for all oractical purposes, to bn

proportional to losses of total mass. Hough (1901)

developed i cegression equation to describe the loss of in

prescribed barns:

N (k~ /ha) = -10.55 + 0.0071 ()

where X is the loss of mass in kg/ha. The ePan I

concentration in the fuel of these stands is aboit 0.7='

(Hough 1992). La a field study not included in Table 12,

Klemmedson at al. (1963) found losses of 30T of mass and 11'

of N after prnscribed burning of ponderosa pine litter. T-

a lab stidy Kni;ht (1966) burned conifer fores- floor

samples in a muffle furnace and found losses of 1Q0 nd 2"?

for mass in! N, respectively, at 300oC and 53 and 61' It

6000C. Tn Lab studies on southeastern oine litter in whiSh

feel samples were simply ignited, Denell and Ralston (197n)

recorded losses of 71% of the mass and 58q of N whi Le',;is

(1975) founi losses of 39 and 56% for mass and N,

respectively. Tr these relatively cool burns most of the '

was lost as linitrogan gas.

Pires may also result in losses o' the so-called ish

elements (Cliyton 1976, Table 12). Rowever, undar nost

circumstances it is not possible to show statistically

significant Losses by preburn and postburn sampling (e.q,

this study and Pichter et al. 1982). The magnitude of these

losses as ocrsentei in much of the literature is lik lv Io

be overestimated. In field studies it is difficult to

sample the ash. When small plots ire burned, and psecia!l?

if samples oF fuel are burned in laboratory settings, thcre

is no opportunity for the particulate matter to settle b-ck

down on the sampled area. Significant amounts of nutrients

in articulate matter are carried away from small ",urns

(Smith and PBwes 1974, Wvans and All-n 1971). Tn nore

extensive burns much of the particulate fraction of smoko

will rttura lirzetlv to the ground. Some of the narticuilate

emissions are deoosited on the canopy ani are returned to

the forest floor by the first rain after the bolrn. Ther is

also a tendency to consider any loss reported in the

literature as rail while gains after burning are iqgored

(e.g. Boerner 1982). For example, the increase of 25! in

the amount of Mg reported by Debans and Conral (1978, Table

12) makes the loss of 16t of K seem of questionable


The ash element most likely to be lost by volatilization

is K, which is volatilized at temperatures above? 500oc. In

what ar3 pr:Sably the two hottest fires listed in Tabl! 1?

(chaparral and the Entiat wildfire) the percietaae of K 0lot

was higher than that of any other ash element. This wae

also true of the lab simulations on heather.

The ash lepositel on the soil surface is subject to

erosion losses om sloped terrain, but not in Soutnh Florila.

Much of the mineral contest of the ash is easily leach-d

into the soil by rainfall (Allen 196i, Lewis 197U,

:hristensen 1977). Recently burned ecosystems may

experience losses to groundwater, however thsse losses s=em

to be very small for coastal plain pinelenls (Boerner an-

Forman 1982, Rizhter et al. 1982). In Long Pine Key th=

high soil organize matter content should giva the soil a hiTh

catioa xcha~7e a paaity and P should be qui'klv immobhli.-~

by the abuindint :a. The rapid recovery o' +t1 vegetation

also acts as a sink for the nutrients released by th" fi-r

(see below).

Althon)h the losses of nutrients (other than N) in the

borns in this study were not statistically significant it is

certain that som? losses did occur. "ow=v-r, evsn if 'n

assumes a loss of 10' of the standing crop oa the ash

elements, tiase losses could be replaced by precipitation

inputs over the period of a burning rotation, 3.'-6 7r.

Annual deposition of N, P, K, Ca, and Eg in bulk

precipitation in northcentral Florida is about 1-1.4, 0.1,

0,3-0.6, 1-1.5, and 0.2 g/ma respectively ('pndry and

Brezonik 199), Riekerk et al. 1979). Depending ipon the

estimate of input by precipitation, this may or may not

replace the substantial loss of V because losses ranqed from

1.4-2.2 g*ma-*yr-i when averaged over the time since= th-

previous fire,

It is to be expected that nutrient losses should more or

less cqual irnuts in ecosystems that have a long history on

frequent burning. In fact, fire may be a mechanism by which

nutrient-poor ecosystems prevent the accumulation of

nutrient capital.

Several agents of biological N-fiiation could aid in

recovery of N lost through burning. Although several

herhaceous laiumns are common (e.g. Cassia spp., Crotalaria

umrill) they are not likelv to fix significant amoilnts

(Pundal 1981). Lvsiloma latisil:l!_ is a legaiino_,s tre-,

but it is not zomaoa enough in pinelands to aid mich

nitrogen. Zimi E 1umila has a blue-green algal endophvt?

that fixes (Lindblad 1984). Eyrica cprifera, an

actinomycete-nodulated shrab, has substantial biomass in

much of Lonj Pine Key (inclulinq site 2) and may fir

appreciable i~ounits of 4. Tn a northern 1lorila slash -);in=

plantation Permar and Fisher (1993) estimated that "

cerifera witi a crown cover of 8" fixed 1.1 a nT*c-2*yr-1.

Nonsymbiotic N fixation by soil microorganisms in forest

ecosystems his often proved to be a rather minor source o' '

relative to arszipitation inputs (rjepkema 1979), Jorgens'n

and Wells (1971) found indications of increase soil

iaetylene relation activity in burned loblollv pine lots

compared to inburned controls. A later stufy by Jor"ensen

(1975) -ii lot support this finding and found 4-fixation

rates of < 0.1 g*m-z*yr-i. Vance at al. (1983) found even

lo er rites (0.01 g*m-2*yr-i) and no effect of long-ter-

prescribed burning in oak-hickory forest. In a orairi?

site, DuBois and Kapustka (1983) measured 0.8 qm-2*vr-1 f

nonsymbioti= T-fixation, about half of which was du- t3

cyanobacterii. Blue-green algae are not found in acid

coastal plaia soils (Jurgensen and Davey 1968) but theyv r?

coamon in ciccumneatral soils (;ranhall and Henriksson l119)

and abound 3i the oolitic limestone of the "iani Rock idgi-.

I ran an acetylene reduction assay on a few samples of rok

front the stuly area and fo:nd measurable rates afteT hr

If incubation. Crude =alculations indicated rates on t+e

order of 0.1 g*m-2*yr-1. It would seem that fixation bv

various taxi plus precipitation inputs over a burning

rotation could easily account for th- losses of N r-corr~l

for the experimental fires.

Pistfire aecoverv

In the !imi Rock Ridge pinelands virtually all recovrv

of understorv species is by vegetativP means rather than hv

seedling reproduction. This is fairly common in ecosystem~

that exnpric= z high fire frequnncies. Abrahanson (193aa)

found very little change in species composition after fire'

in several pLn?-1ominated plant coamunities in southcuntral

"lorida. Bo2rne' (1981) also note relatively little chaenr

in speci-s c=apo3ition in burned areas of the vew Jorse-

Pine 9arrens,


This is La contrast to situations, often whsre fires are

of lower frequency but higher intensity, in whi-h there i a

distinctive )ostburn flora. The large number of herbaceoii

annuals that appear after chaparral fires is a Iramatic

example (Christensen and Iuller 1975). The degree to which

species changes result from burning depends in part on the

severity of the fire. In jack pine communities, fires

during the Iry summer months consume the upnse organic

layers of tie soil, whereas fires during the moister snrimn

do not. Sia3e species with shallow regenerativ- organs ~a

eliminated by ground fires, numerous postfire disturbanr.c

species like Fpilobinm (firewesd) become established after

summer fires (Ohmann and Grigal 1981).

Even though recovery in the rock ridq3 pin-lands is

vegetative, there are differences among the different growth

forms in the rat3 of recovery. Herbs and palms renrow v=rv

quickly, reaching their preburn dry masses within 1 yr. Tho

monocots (jcaminaid herbs and palms) and ferns havy

protected maristems that are seldom damaaed by Fires.

grasses are sell known for their rapid regrowth after top

removal (Iiltan and Lewis 1962, Daubenmire 1968). The

recovery of licots (whose buds are usually killed) rcuirrs

the activation of previously inactive maristems or the

production 3f aiventitious buds, whether thny bp on

rhizomes, roats, or the base of stems; woody dicots take

longer than iarbaieous dicots to show regrowth. eowver, a


study by Hounh (1965) in =eorgia found that Serenoa repnen

only reached 55- of its preburn mass 1 yr after burning

whereas 11 x labra, a hardwood, returned to 70q of its

preburn leveL. This is somewhat anomalous because the

preburn vegetation had bean unburned for 15 vr and there

should have been considerable stem biomass. S. reoons was

found to have greater cover 1 yr after a burn 'thn heforp

the born in i study by Abrahamson (198qb). This agrees more

closely with my findings in the rock ridge Dinelands.

Iv 1 yr IEtae the experimental burns, 130-190 7/m? of

regrowth vegetation appeared in the burned plots. This is a

fairly large amount in relation to some other postf-irn

recoveries. One year after a prescribed burn in the nw

Jersey Pine Sirrens the regrowth of herbs, shrubs, and oak

sprouts was 113 1/a1 (Boeraer 1991). In jack pin- stands

baraed by wilifires-in Mianesota the vegetation recovery

ranged from 21-131 g/a2 after one complete growing season

(Ohmann and 3rigal 1981).

The raeovrcy of photosynthetic capacity is ev=n more

rapid than the recovery of biomass because hardwoods orolnce

leaves more rapillv than support tissues. In the four burns

in Long Pine Key the recovery of ohotosynthetic tiss' s (sum

of herbs, palms, and hardwood leaves) was 59-937 of th2

prohurn amount at the end of 1 yr. Th, first ve:r recovery

of hardwood leaves alone ranged from 32-68" of the initial

quantities. The rapid recovery of photosynthetic surface is

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs