<%BANNER%>

Method Development to Characterize Nutrient Release Patterns of Enhanced-Efficiency Fertilizers

Permanent Link: http://ufdc.ufl.edu/UFE0042750/00001

Material Information

Title: Method Development to Characterize Nutrient Release Patterns of Enhanced-Efficiency Fertilizers
Physical Description: 1 online resource (154 p.)
Language: english
Creator: MEDINA,LIDYA CAROLINA
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ACCELERATED -- CHARACTERIZATION -- CONTROLLED -- CORRELATION -- EFFICIENCY -- ENHANCED -- FERTILIZER -- INCUBATION -- LEACHING -- NUTRIENT -- RELEASE
Soil and Water Science -- Dissertations, Academic -- UF
Genre: Soil and Water Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Several technologies have been proposed to characterize the nutrient release patterns of enhanced efficiency fertilizers (EEFs) during the last few decades. These technologies have been developed mainly by manufacturers and are product-specific based on the regulation and analysis of each EEF product. Despite previous efforts to characterize EEF materials, no official method exists to assess their nutrient release patterns. However, the increased production and distribution of EEFs in specialty and non-specialty markets requires an appropriate method to verify nutrient release claims and material performance. A soil incubation-column leaching procedure was evaluated to determine its suitability as a standard method to estimate nitrogen (N) release patterns of EEFs during 180 d. The influence of three sand/soil ratios, three incubation temperatures, and four soil types on method behavior was assessed using eight EEFs. In general, the highest sand/soil ratio had an effect on the N release rate of the EEFs that varied depending on the type of fertilizer, but was more marked for the slow-release fertilizers (SRF). Temperature had the greatest influence on EEF N release rates. For controlled-release fertilizers (CRF), the initial N release rates and the percentage N released per day increased as temperature increased. For SRFs, raising the temperature from 25 to 35?C increased initial N release rate and the total cumulative N released, and almost doubled the percentage N released per day. The percentage N released per day from all EEFs generally increased as the texture of the soil changed from sandy to loamy (Iowa>California>Pennsylvania>Florida). A series of experiments were conducted to evaluate the effect of temperature, fertilizer sample size, and extraction time on the performance of a 74-h accelerated lab extraction method to measure EEF nutrient release profile. Temperature was the only factor that influenced nutrient release rate, with a highly marked effect for phosphorus (P) and to a lesser extent for N and potassium (K). Based on the results, the optimal extraction temperature set was: Extraction #1- 2:00 hr.@ 25?C; Extraction #2- 2:00 hr.@ 50?C; Extraction #3- 20:00 hr. @ 55?C; and Extraction #4- 50:00 hr. @ 60?C. Ruggedness of the method was tested by evaluating the effect of small changes in seven selected factors on method behavior using a fractional multifactorial design. Overall, the method showed ruggedness for measuring N release rates of coated EEFs. Non-linear regression was used to establish a correlation between the data generated from both methods, and to develop a model that can predict the 180-d N release curve for a specific EEF product based on the data from the accelerated lab extraction method. Based on the R2 > 0.90 obtained for most EEF materials, results indicated that the data generated from the 74-h accelerated lab extraction method could be used to predict N release from the selected EEFs during 180 d, including those fertilizers that require biological activity for N release.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by LIDYA CAROLINA MEDINA.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Obreza, Thomas A.
Local: Co-adviser: Sartain, Jerry B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042750:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042750/00001

Material Information

Title: Method Development to Characterize Nutrient Release Patterns of Enhanced-Efficiency Fertilizers
Physical Description: 1 online resource (154 p.)
Language: english
Creator: MEDINA,LIDYA CAROLINA
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ACCELERATED -- CHARACTERIZATION -- CONTROLLED -- CORRELATION -- EFFICIENCY -- ENHANCED -- FERTILIZER -- INCUBATION -- LEACHING -- NUTRIENT -- RELEASE
Soil and Water Science -- Dissertations, Academic -- UF
Genre: Soil and Water Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Several technologies have been proposed to characterize the nutrient release patterns of enhanced efficiency fertilizers (EEFs) during the last few decades. These technologies have been developed mainly by manufacturers and are product-specific based on the regulation and analysis of each EEF product. Despite previous efforts to characterize EEF materials, no official method exists to assess their nutrient release patterns. However, the increased production and distribution of EEFs in specialty and non-specialty markets requires an appropriate method to verify nutrient release claims and material performance. A soil incubation-column leaching procedure was evaluated to determine its suitability as a standard method to estimate nitrogen (N) release patterns of EEFs during 180 d. The influence of three sand/soil ratios, three incubation temperatures, and four soil types on method behavior was assessed using eight EEFs. In general, the highest sand/soil ratio had an effect on the N release rate of the EEFs that varied depending on the type of fertilizer, but was more marked for the slow-release fertilizers (SRF). Temperature had the greatest influence on EEF N release rates. For controlled-release fertilizers (CRF), the initial N release rates and the percentage N released per day increased as temperature increased. For SRFs, raising the temperature from 25 to 35?C increased initial N release rate and the total cumulative N released, and almost doubled the percentage N released per day. The percentage N released per day from all EEFs generally increased as the texture of the soil changed from sandy to loamy (Iowa>California>Pennsylvania>Florida). A series of experiments were conducted to evaluate the effect of temperature, fertilizer sample size, and extraction time on the performance of a 74-h accelerated lab extraction method to measure EEF nutrient release profile. Temperature was the only factor that influenced nutrient release rate, with a highly marked effect for phosphorus (P) and to a lesser extent for N and potassium (K). Based on the results, the optimal extraction temperature set was: Extraction #1- 2:00 hr.@ 25?C; Extraction #2- 2:00 hr.@ 50?C; Extraction #3- 20:00 hr. @ 55?C; and Extraction #4- 50:00 hr. @ 60?C. Ruggedness of the method was tested by evaluating the effect of small changes in seven selected factors on method behavior using a fractional multifactorial design. Overall, the method showed ruggedness for measuring N release rates of coated EEFs. Non-linear regression was used to establish a correlation between the data generated from both methods, and to develop a model that can predict the 180-d N release curve for a specific EEF product based on the data from the accelerated lab extraction method. Based on the R2 > 0.90 obtained for most EEF materials, results indicated that the data generated from the 74-h accelerated lab extraction method could be used to predict N release from the selected EEFs during 180 d, including those fertilizers that require biological activity for N release.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by LIDYA CAROLINA MEDINA.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Obreza, Thomas A.
Local: Co-adviser: Sartain, Jerry B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042750:00001


This item has the following downloads:


Full Text

PAGE 1

1 METHOD DEVELOPMENT TO CHARACTERIZ E NUTRIENT RELEASE PATTERNS OF ENHANCED EFFICIENCY FERTILIZERS By CAROLINA MEDINA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2 011

PAGE 2

2 2011 Carolina Medina

PAGE 3

3 To my husband and my family

PAGE 4

4 ACKNOWLEDGMENTS I would like to express my most sincere gratitude and appreciation to Dr. Thomas Obreza Dr. Jerry Sartain and Mr. William Hall for their guidance assistance and knowledge throughout my doctorate program Their support has helped me develop both professionally and personally I am thankful to Dr. Ann Wilkie and Dr. Jason Kruse for serving in my supervisory committee and provid ing me with their comments for my research project. Further, my special thanks go to Greg Means and the staff of the Soil Fertility and Turfgrass Nutrition Laboratory for their friendly assistance through laboratory and greenhouse tasks for my research. I am also thankful to Emily Leary for her invaluable assistance with statistics. Finally, I would like to thank my husband family and friends for their constant love, support and help which inspired me to complete this de gree

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ ........ 10 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 I NTRODUCTION AND LITERATURE REVIEW ................................ ..................... 15 Enhanced Efficien cy Fertilizers ................................ ................................ ............... 16 Types of Enhanced Efficiency Fertilizers ................................ ................................ 17 Factors Affecting Nutrient Release of EEF ................................ ............................. 20 Methods to Estimate Nutrient Release Patterns of EEF ................................ ......... 25 2 EVALUATION OF A SOIL INCUBATION METHOD TO CHARACTERIZE NITROGEN RELEASE PATTERNS OF ENHANCED EFFICIENCY FERTILIZERS ................................ ................................ ................................ ......... 29 Materials and Methods ................................ ................................ ............................ 31 Study #1: Sand/Soil Ratio Effect ................................ ................................ ...... 33 Study #2: Incubation Temperature Effect ................................ ......................... 33 Study #3: Soil Type Effect ................................ ................................ ................ 34 Statistical Analysis ................................ ................................ ............................ 35 Results and Discussion ................................ ................................ ........................... 35 Study #1: Sand/Soil Ratio Effect ................................ ................................ ...... 35 Study #2: Incubation Temperature Effect ................................ ......................... 41 Study #3: Soil Type Effect ................................ ................................ ................ 44 3 OPTIMIZATION AND VALIDATION OF AN ACCELERATED LAB EXTRACTION METHOD TO ESTIMATE NITROGEN RELEASE PATTERNS OF ENHANCED EFFICIENCY FERTILIZERS ................................ ....................... 69 Materials and Methods ................................ ................................ ............................ 71 Standard Accelerated Lab Extraction Method ................................ .................. 71 Optimization of the Accelerated Lab Extraction Method ................................ ... 74 Preliminary temperature study ................................ ................................ ... 74 Method optimization studies ................................ ................................ ....... 75 Statistical analysis ................................ ................................ ...................... 76 Ruggedness Testing of the Accelerated Lab Extraction Method ...................... 76 Results and Discussion ................................ ................................ ........................... 79 Optimization of the Accelerated Lab Extraction Method ................................ ... 79

PAGE 6

6 Ruggedness Testing of the Accelerated Lab Extraction Method ...................... 83 4 STATISTICAL CORRELATION OF THE SOIL INCUBATION AND THE ACCELERATED LAB EXTRACTION METHODS TO ESTIMATE NITROGEN RELEASE RATES OF ENHANCED EFFICIENCY FERTILIZERS ....................... 105 Materials and Methods ................................ ................................ .......................... 106 Nonlinear Regression ................................ ................................ ..................... 106 Prediction Models ................................ ................................ ........................... 107 Data issues ................................ ................................ .............................. 108 Zero extraction values ................................ ................................ .............. 110 Two step nonlinear regression method with grouping .............................. 111 Principal component analysis with grouping ................................ ............ 112 Two step method based on PCA without grouping ................................ .. 113 Results and Discussion ................................ ................................ ......................... 114 Final Prediction Model ................................ ................................ .................... 116 Error Estimation ................................ ................................ .............................. 117 5 CONCLUSIONS ................................ ................................ ................................ ... 141 Long Term Soil Incubation M ethod ................................ ................................ ....... 141 Short Term Accelerated Laboratory Extraction Method ................................ ........ 143 Correlation of the Soil Incubation and the Accelerated Extractio n Methods ......... 145 LIST OF REFERENCES ................................ ................................ ............................. 147 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 154

PAGE 7

7 LIST OF TABLES Table page 2 1 Enhanced efficiency fertilizer specifications. ................................ ...................... 54 2 2 Physical and chemic al characteristics of the soils ................................ .............. 54 2 3 Regression analysis of estimated N release rate from different CRFs using the soil incubation met hod with three sand/soil ratios ................................ ......... 55 2 4 Regress ion analysis of estimated N release rate from different SRFs using the soil incubation method with three sand/soil ratios. ................................ ........ 57 2 5 Effect of sand/soil ratio on total N released from five CRFs wi th time ................ 58 2 6 Effect of sand/soil ratio on total N released from three SRFs with time .............. 59 2 7 Regression analysis of estimated N r elease rate from different CRFs using the soil incubation method at three incubation temperatures. ............................ 60 2 7 Continued ................................ ................................ ................................ ........... 61 2 8 Regress ion analysis of estimated N release rate from different SRFs using the soil incubation method at three incubation temperatures. ............................ 62 2 9 Effect of temperature on total N released from five CRFs w ith time ................... 63 2 10 Effect of temperature on total N released from three SRFs with time ................. 64 2 11 Regression analysis of estimated N re lease rate from five CRFs using the soil incubation method with four different soils. ................................ .................. 65 2 11 Continued. ................................ ................................ ................................ .......... 66 2 12 Effect of soil ty pe on total N released from five CRFs with time ......................... 67 2 12 Continued ................................ ................................ ................................ ........... 68 3 1 Sequence of extraction times and temperatures for the sta ndard accelerated lab extraction method. ................................ ................................ ........................ 94 3 2 Sequence of extraction times and temperatures used in the preliminary temperature study. ................................ ................................ .............................. 94 3 3 Description of the enhanced efficiency fertilizer u sed in the optimization studies ................................ ................................ ................................ ................ 94

PAGE 8

8 3 4 Factors and their levels for ruggedness testing of the accelerated lab extraction method. ................................ ................................ .............................. 95 3 5 Enhanced efficiency fertilizers used in the ruggedness testin g of the accelerated lab method ................................ ................................ ...................... 95 3 6 Experimental design of ............................ 96 3 7 Parameters used in the statistical interpretation of the logarithm ................................ ................................ ................................ ............. 96 3 8 Rankits used to draw half normal probability plots using eight experiments ....... 96 3 9 Effect of four temperature sequences on total cumulative %N released from two EEFs using the standard ac celerated lab extraction method ....................... 97 3 10 Effect of temperature, fertilizer sample size, and extraction time on total N released from four EEFs using the ac celerated lab extraction method ............... 98 3 11 Effect of temperature, fertilizer sample size, and extraction time on total P released from two EEFs using the ac celerated lab extraction method ............... 99 3 12 Effect of temperature, fertilizer sample size, and extraction time on total K released from two EEFs using the accelerated lab extraction method. ............ 100 3 13 Results of e ight ruggedness experiments on the cumulative %N released for five EEFs using the ac celerated lab extraction method ................................ .... 101 3 14 Results of eight ruggedness experiments on the cumulative %P release d for two EEFs using the ac celerated lab extraction method ................................ .... 102 3 15 Results of eight ruggedness experiments on the cumulative %K released for two EEFs using the ac celerated lab extraction metho d ................................ .... 102 3 16 Effects of seven factors on the cumulative %N released for five EEFs using the experimental error ................................ .... 103 3 17 Effects of seven factors on the cumulative %P released for two EEFs using the experimental error ................................ .... 104 3 18 Effects of seven factors on the cum ulative %K released for two EEFs using the experimental error ................................ .... 104 4 1 Statistics from ANOVA of fitted N release curve coefficients by group ............. 136 4 2 Enhanced efficiency fertilizers used for the statistical c orrelation of the methodologies ................................ ................................ ................................ .. 137

PAGE 9

9 4 3 Regression coefficients and R 2 values for the fitted soil incubation N release data for twelve EEFs ................................ ................................ ........................ 138 4 4 L eave one out cross validation results for the two step nonlinear r egression method and the PCA with grouping method for twelve EE Fs ........................... 139 4 5 Regression coefficients and R 2 values for the predicted N release curves using the final selected method ( two step nonlinear regression) for twelve EEFs. ................................ ................................ ................................ ................ 140

PAGE 10

10 LIST OF FIGURES Figure page 2 1 Incubation lysimeters. ................................ ................................ ......................... 45 2 2 Nitrogen release curves for SCU, resin coated NPK Poly S, and RLCU as affected by soil sample size using a soil incubation method ............................... 46 2 3 Nitrogen release curves for polyolefin coated NPK, IBDU, ureaform and biosolids as affected by soil samp le size using a soil incubation method ........... 47 2 4 Effect of N source on leachate pH using the soil incubation method, A= low soil sample size (1755 g sand: 45 g soil); B=standard soil sample size ( 1710 g sand: 90 g soil); C=high soil sample size (1620 g sand: 180 g soil) ................ 48 2 5 Effect of N source on leachate pH using the soil incubation method, A= standard temperature (21C); B=medium te mperature (2 5C); C=high temperature (35C) ................................ ................................ ............................. 49 2 6 Nitrogen release curves for SCU, resin coated NPK, Poly S, and RLCU as affected by incubation temperature using a soil incubation method. .................. 50 2 7 Nitrogen release curves for polyolefin coated NPK, IBDU, ureaform, and biosolids as affected by incubation temperature using a soil incubation method. ................................ ................................ ................................ .............. 51 2 8 Effect of N source on leachate pH using the soil incubation method. A=Florida soil; B=California soil; C=Pen nsylvania soil, and D=Iowa soil ............ 52 2 9 Nitrogen relea se curves for SCU, resin coated NPK, Poly S, Polyolefin coated NPK, and RLCU as affected by soil type using a soil incubation method. ................................ ................................ ................................ .............. 53 3 1 Extraction apparatus with eight jacketed chromatogra phy columns. .................. 87 3 2 Schematic diagram of water manifold used in the extraction apparatus. ............ 87 3 3 Schematic diagram of the extracti on phase. ................................ ....................... 88 3 4 Schematic diagram of the collection phase ................................ ........................ 88 3 5 Nitrogen released from POC NPK Type A and Type B in 168 hr as af fected by four e xtraction temperature sequences ................................ ......................... 89 3 6 Effect of temperature, fertilizer sample size, and extraction time on N released from ureaform, PCU, PC NPK Type A, and PC NPK Type B in 7 4 hr compared with the standard accelerated lab extraction method. ........................ 90

PAGE 11

11 3 7 Release of N, P, and K from two EEFs in 74 hr using the accelerated lab extraction method (a: PC NPK Type A at standard temperature; b: PC NPK Type A at medium temperature; c: PC NPK Type B at standard temperature; d: PC NPK Type B at medium temperature). ................................ ...................... 91 3 8 Half normal probability plot for seven effects on percentage N released from RLCU with identification of the critical effects ME and SME. .............................. 92 3 9 Half normal probability plot for seven effects on percentage N released from IBDU with identifica tion of the critical effects ME and SME. ............................... 92 3 10 Half normal probability plot for seven effects on percentage K released from resin coated NPK with identification of the critical effects ME and SME ............. 93 4 1 Scree plots for four groups of EEFs ................................ ................................ 120 4 2 Predicted N release curves using the two step nonlinear regression method with 2, 3, and 4 extraction values for ureaform, methylene urea A, RLC urea, and SCU ................................ ................................ ................................ ........... 121 4 3 Predicted N release curves using the two step nonlinear regression method with 2, 3, and 4 extracti on values for IBDU, resin coated NPK, blend A, and Poly S ................................ ................................ ................................ ............... 122 4 4 Predicted N release curves using the two step nonlinear regression method with two, three, and four extraction values for polyo lefin coated NPK, methylene urea A methylene urea C, and blend B ................................ .......... 123 4 5 Predicted N release curves using the PCA with grouping method with two and three principle components for ureaform, met h ylene urea A, RLC urea, and SCU ................................ ................................ ................................ ........... 124 4 6 Predicted N release curves using the PCA with grouping method with two and three principle components for IBDU, resin coated NPK, blend A, and Poly S ................................ ................................ ................................ ............... 125 4 7 Predicted N release curves using the PCA with grouping method with two and three principle components for polyolefin coated NPK, methylene urea A, methylene urea C, and blend B. ................................ ................................ ... 126 4 8 Predicted N release curves using the three selected prediction methods for ureaform, methylene urea A, RLC urea, and SCU (Model 1: two step nonlinear regression method; Model 2: PCA with groupin g method; and Model 3: PCA like non stratified method). ................................ ........................ 127 4 9 Predicted N release curves using the three selected prediction methods for IBDU, resin coated NPK, blend A, and Poly S (Model 1: two step nonlinear

PAGE 12

12 regression method; Model 2: PCA with grouping method; and Model 3: PCA like non stratified method) ................................ ................................ ................ 128 4 10 Predicted N release curves using the three selected prediction met hods for polyolefin coated NPK, methylene urea A, methylene urea C, and blend B (Model 1: two step nonlinear regression method; Model 2: PCA with grouping method; and Model 3: P CA like non stratified method) ................................ .... 129 4 11 Predicted N release curves using the final prediction method (two step nonlinear regression) for ureaform, met hylene urea A, RLC urea, and SCU .... 130 4 12 Predicted N relea se curves using the final prediction method (two step nonlinear regression) for IBDU, resin coated NPK, blend A, and Poly S. ......... 131 4 13 Predicted N release curves using the final prediction metho d (two step nonlinear regression) for polyolefin coated NPK, methylene urea A, methylene urea C, and blend B. ................................ ................................ ....... 132 4 14 Predicted N release curves using the final method with the corresponding 9 0% prediction intervals for ureaform, methylene urea A, RLC urea, and SCU. ................................ ................................ ................................ ................. 133 4 15 Predicted N release curves using the final method with the corresponding 90% prediction intervals for IBDU, r esin coated NPK, blend A, and Pol y S ..... 134 4 16 Predicted N release curves using the final method with the corresponding 90% prediction intervals for polyolefin coated NPK, methylene urea A, meth ylene urea C, and blend B. ................................ ................................ ....... 135

PAGE 13

13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MET HOD DEVELOPMENT TO CHARACTERIZ E NUTRIENT RELEASE PATTERNS O F ENHANCED EFFICIENCY FERTILIZERS By Carolina Medina May 2011 Chair: Thomas Obreza Cochair: Jerry Sartain Major: Soil and Water Science Several technologies have been proposed to characterize th e nutrient release patterns of enhanced efficiency fertilizers ( EEFs ) during the last few decades. These technologies have been developed mainly by manufacturers and are product specific based on the regulation and analysis of each EEF product Despite pre vious efforts to characterize EEF materials, no official method exists t o asses s their nutrient release patterns. However, the increased production and distribution of EEFs in specialty and non specialty markets requires an appropriate method to verify nut rient release claims and material p erformance. A soil incubation column leaching procedure was evaluated to determine its suitability as a standard method to estimate nitrogen ( N ) release patterns of EEFs during 180 d The influence of three sand/soil rati o s, three incubation temperature s and four soil types on method behavior was assessed using eight EEFs In general, the high est sand/soil ratio increased the N release rate of the EEFs but this effect was more marked for the slow release fertilizers (SRF ). Temperature had the greatest influence on EEF N release rates. For controlled release fertilizers (CRF), the initial N

PAGE 14

14 release rates and the percentage N release d per day increased as temperature increased. For SRFs, raising the temperature from 25 to 3 5C increased initial N release rate and the total cumulative N released, and almost doubled the percentage N released per day. The percentage N release d per day from all EEFs generally increased as the t exture of the soil changed from sandy to loamy ( Iowa >California>Pennsylvania>Florida ) A series of experiments were conducted to evaluate the effect of temperature, fertilizer sample size, and extraction time on the performance of a 74 h accelerated lab extraction method to measure EEF nutrient release prof ile. Temperature was the only factor that influenced nutrient release rate with a highly marked effect for phosphorus ( P ) and to a lesser extent for N and potassium ( K ) Based on the results, the optimal extraction temperature set was: Extraction #1 2:00 hr.@ 25C; Extraction #2 2:00 hr.@ 50C; Extraction #3 20:00 hr. @ 55C; and Extraction #4 50:00 hr. @ 60C. Ruggedness of the method was test ed by evaluating the effect of small changes in seven selected factors on method behavior using a fractional m ultifactorial design. Overall, th e method showed rugged ness for measuring N release rates of coated EEFs. Non linear regression w as used to establish a correlation between the data generated from both method s and to develop a model that can predict the 1 80 d N release curve for a specific EEF product based on the data from the accelerated lab extraction method Based on the R 2 > 0.90 obtained for most EEF materials, results indicated that the data generated from the 74 h accelerated lab extraction method could be used to predict N release from the selected EEFs during 180 d including those fertilizers that require biological activity for N release.

PAGE 15

15 CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW The u se of c hemical fertilizers is the main way supplemental n utrient s are provided to crop s Generally, at least 30 to 50% of crop yield is attributed to commercial fertilizer nutrient inputs (Stewart et al., 2005). The intensification of agriculture mainly due to population growth and agriculture technological adv ancements has accelerated fertilizer use even more during the last century. World fertilizer consumption is expected to grow about 1.7% annually from 2007/2008 to 2011/2012, which is an increment equivalent to about 15 million tons (FAO 2008). However, c r op recovery of applied inorganic fertilizer is estimated to be 50% or lower for N less than 10% for P and close to 40% for K (Baligar et al., 2001). Most fertilizers are readily water soluble when applied to the soil Subsequently they undergo numerous c omplex interactions between plant roots, soil microorganisms, chemical reactions and loss pathways. Th e discrepancy between increased fertilizer consumption and relatively low nutrient use efficiency by crops is due to significant losses of nutrients to ai r groundwater and surface water or fixation of nutrients by soil The main losses of N fertilizer applied to soil are volatilization of ammonia (NH 3 ), leaching of nitrate (NO 3 ), gaseous losses of N 2 and N 2 O after denitrification of NO 3 runoff and er osion ( Ei c khout et al., 2006 ). These losses result not only in the contamination of the environment with the degradation of soil, air, and water quality but also in negative economic effects for farmers. Enhanced efficiency fertilizer gradually release s nu trients to match nutrient release patterns with crop demand and therefore improve s nutrient recovery by plants. Because of prolonged nutrient availability, EEFs can potential ly minimize negative environmental effects of fertilization that are largely due t o the high solubility of N compounds applied

PAGE 16

16 to the soil ( L ubkowski and Grzmil, 2007). The market share of EEFs amounts to only 0.47% of world mineral fertilizer consumption, even taking into account the large increase in production capacity since 2005. Al though the EEF market is almost negligible, its world consumption increased by 45% in 2004/05 compared with 1995/96, indicating a fast growing market that is mainly due to the additional production capacity of EEFs and their increased use on agricultural c rops (Trenkel 2010). A wide variety of EEFs are currently produced and distributed in United States, Canada, China, Japan, Europe and Israel for specialty and non specialty markets. However, there is no official method to verify nutrient release patterns of these fertilizers. Traditionally, each technology has been addressed by its own manufacturer in terms of nutrient release characterization and performance. However, as the use of EEFs increase s an individual approach towards label claim verification a nd nutrient release performance is no longer sufficient There is a need to develop a standard method that can be used to estimate nutrient release patterns of a broad range of EEF products. Enhanced Efficiency Fertilizers The term enhanced efficiency fer tilizer refer s to fertilizer products with properties that allow increased plant uptake and minimize the potential of nutrient loss to the environment (e.g., leaching, volatilization and runoff ) EEFs include slow and controlled release fertilizers that provide extended nutrient suppl y to a crop. According to the American Association of Plant Food Control Officials (APPFCO), which officially defines and regulates these terms, there is no differen ce between slow release and controlled release fertilizers. Slow or controlled release fertilizer i s officially defined by APPFCO as a fertilizer containing a plant nutrient in a form that delays its availability for plant uptake and use after application, or that extends its availability to the plant

PAGE 17

17 significantl y longer than a reference rapidly available nutrient fertilizer such as ammonium nitrate urea, ammonium phosphate or potassium chloride. Such delay of initial availability or extended time of continued availability occur s by a variety of mechanisms. Th ese include controlled water solubility of the material by semi permeable coatings, occlusion, protein materials, or other chemical forms, by slow hydrolysis of water soluble low molecular weight compounds, or by other unknown means (APPFCO Official Public ation 57). According to Shaviv (2001), the difference between slow and controlled release fertilizer products lies in the release mechanisms and environmental factors that affect their delayed or extended availability of nutrients. Slow release fertilizer s are manufactured to release their nutrients with time at a slow rate as they undergo different decomposition processes in the soil, such that their rate, pattern and release period depend on soil and climatic conditions. In contrast, controlled release fertilizers are manufactured to gradually release nutrients at a rate that matches plant nutrient demand such that the rate, pattern and duration of release are well known and controlled. Types of Enhanced Efficiency Fertilizers According to Goertz (1993) EEF s can be classified into two types: (1) urea aldehyde condensation products that are produced by the chemical reaction of water soluble urea or ammonia compounds to produce N fertilizers with more complex molecular structures that have limited water s olubility and slowly decompose in the soil, and (2) coated fertilizer products that are made by coating or encapsulating a water soluble fertilizer with a water insoluble barrier that controls the access of water to the fertilizer, and thus limits its diss olution rate.

PAGE 18

18 Urea aldehyde condensation products can be further divided into urea formaldehyde (UF) reaction products that decompose biologically in the soil and other urea aldehyde compounds that decompose mainly chemically such as isobutylidene diurea (IBDU). The reaction of urea with formaldehyde results in a distribution of methylene urea (MU) polymers of varying chain lengths (molecular weights) and consequently varying water solubility. Based on the degree of polymerization, UF products can be divi ded in to three classes: 1 ) ureaform ( the least water soluble class ) ; 2 ) methylene ureas ( intermediate chain length MU polymers ) ; and 3 ) methylene diurea/dimethylene triurea com binations ( the shortest chain MU oligomers and the most water soluble ) (Sartain and Kruse 2001). The polymer chain length affects the rate of N release from UF reaction products T he longer the MU polymer, the longer it takes for N to release. Release of N from UF products requires dissolution that occurs slowly because of their low solubility, followed by microbial decomposition or hydrolysis in the soil. Biodegradation of UF occurs mainly by microbial decomposition at the carbon bond in the MU polymers rather than by hydrolysis. IBDU is produced by the reaction of urea and isobutyra ldehyde, resulting in a single water insoluble oligomer. Once dissolution of IBDU takes place, N becomes available to plants through hydrol y sis that proceeds rapidly in the soil with regeneration of the original reactants (urea and isobutyraldehyde). The r ate of dissolution of IBDU is determined by particle size, with smaller particles dissolving more rapidly. Once urea is released from these N reaction products, ammonification and nitrification proceed rapidly in the soil to convert the urea N into nitrate N for plant uptake (Allen, 1984).

PAGE 19

19 Coated fertilizers refer to products possessing a readily water soluble fertilizer granule surrounded by an impermeable or semi permeable coating that controls the rate of nutrient diffusion. There are three main types o f coating materials: sulfur (S) coating, polymeric coating, and a hybrid material composed of a multilayer coating of S and polymer (Sartain and Kruse 2001). Sulfur coated fertilizer is produced by precisely applying three coating materials (molten S a h ydrocarbon/oil sealant, and a flow conditioner) to heated soluble fertilizer granules. The release of nutrients is determined by imperfections produced by cracks or incomplete coverage of the granule with S and by microbial decomposition of the wax sealan t surrounding the S coating. Once water penetrates the coating through imperfections and/or pinholes produced by microbes, the fertilizer core is dissolved and a rapid release of the nutrient follow s (Oertli, 1980). Polymer coated fertilizers (PCF) release their nutrients by diffusion through a semi permeable coating C onsequently, the composition, quality and thickness of the coating determine the rate of nutrient release. Polymer coatings are classified as either thermoset resins such as alkyd type resins and polyurethane like coatings (also known as reactive layer coating s ), or thermoplastic resins such as polyethylene and polyolefin coatings (Shaviv, 2001). The diffusion release of PCF consists of three stages: 1) i nitial stage with practically no releas e (lag period) during which water vapor enters t h rough the coating due to the vapor pressure gradient across the coating followed by dissol ution of a small fraction of the solid fertilizer core; 2) s tage of constant release rate that starts when a critic al volume of saturated solution accumulates inside the granule and remains constant as long as the saturated solution in the granule is equilibrated with the non dissolved solid fertilizer The dissolved nutrients diffuse out of the granule

PAGE 20

20 due to a concen tration gradient across the coating or are pressed out through mass flow due to the high internal hydrostatic pressure ; and 3) s tage of gradual decay of the release rate that occurs when the solid fertilizers in the core are dissolved and the concentration of the internal solution decreases because of the continuing concomitant fluxes of nutrient release out and water flow into the granule, which in turn, decreases the diffusion release rate ( Shaviv et al., 2003 a ). Polymer sulfur coated (Poly S) fertilizers are modified products in which the S coated fertilizer receives a secondary thin layer of polymer coating to provide a membrane that control s the rate of water and nutrient diffusion into and out of the fertilizer particle. This dual coating permits unifo rm nutrient release approaching PCF performance and provides a positive cost/benefit value compared with products possessing singular coatings of S or polymer. Nutrients from Poly S are released by diffusion of water vapor through the polymeric membrane la yer followed by the subsequent penetration of the water through defects in the S coating by capillary action, solubilizing the fertilizer core. The dissolved nutrients then exit to the soil in reverse sequence (Goertz, 1993). Factors Affecting Nutrient Rel ease of EEF Results from numerous field trials show that environmental factors such as temperature, moisture, and pH affect the hydrolysis and decomposition rates of urea aldehyde condensation products (Jahns et al., 2003). As ureaform is biologically degr aded by microorganisms, its degradation rate is mainly influenced by the innate biological properties of the soil, temperature, and to a limited extent moisture, whereas, pH and particle size have no pronounced effect (Alexander and Helm, 1990). The import ance of soil biological activity level was demonstrated by measuring the mineralization rate of ureaform in three different soils r esult ing in faster mineralization

PAGE 21

21 in the soil with a higher rate of biological activity compared with soils having a compara tively low er rate The mineralization rate of ureaform also increase d greatly as temperature increased from 12 to 32C, reaching a maximum at about 32C and decreasing significantly at temperatures <10C (Alexander and Helm, 1990). Likewise, a study by Koi vunen and Horwath (2004) on the effect of soil management history and temperature on the mineralization of MU showed that a 10 C increase in temperature doubled the mineralization rate. Furthermore, MU degradation rates increased more with increasing tempe rature in soil with lower microbial activity indicating that the degradation of MU was primarily limited by the low activity of the microbes. Dissolution is the rate limiting step in the conversion of IBDU to plant available N forms. While the subsequent hydrolysis of urea to ammonium and its oxidation to nitrate is microbially mediated, the dissolution and hydrolysis of IBDU to urea is completely independent of microbial activity. The rate of dissolution and hydrolysis of IBDU is accelerated by smaller gr anules, higher soil moisture, lower soil pH, and higher temperature (Hughes, 1976; Lunt and Clark, 1969). The nutrient release rate of coated fertilizers is directly affected by coating thickness, coating quality, temperature and to a less er extent moistu re. Sulfur coated urea is the only coated fertilizer that depends directly on microbial activity (for degradation of coating) to release N. Soil temperature and moisture affect microbial activity and consequently the degradation rate of the S coating as we ll as the rate at which urea diffuses out of the granules (Jarrel and Boersma, 1979). Furthermore, greenhouse and laboratory soil studies conducted by Allen et al. (1971) show ed that the rate of SCU dissolution increased greatly with higher temperatures du ring cropping or

PAGE 22

22 incubation and decreased due to thicker S coating and surface application compared with mixing with the soil. SCU particles with thicker S coatings have fewer imperfections than particles with lighter coatings, thus the rate of N release i s slower and vice versa. Temperature is the major factor affecting the rate of nutrient release from PCF. Ahmed et al. ( 1963 ) reported that r ate of fertilizer release from a coated material was directly related to temperature but it was generally indepen dent of the presence o f cr op s, soil texture, and soil moisture between 25 and 100% field moisture capacity. Kochba et al. ( 1990 ) determine d that the main effect of temperature is on water vapor pressure. Since vapor pressure is an exponential function of t emperature, the change of nutrient release rate with temperature is expected to be exponential as well, i.e., the rate of release increases more steeply at higher temperatures. It was also suggested that the properties of the coating materials change with temperature. Oertli and Lunt (1962) speculated that the resin coating of a PCF could possibly soften with increasing temperature and expand slightly because of the internal hydrostatic pressure, causing the pore diameter to increase thus facilitating the diffusion of nutrients. Likewise, Gandeza et al. ( 1991 ) concluded that an increase in temperature increases the moisture permeability of the coating of a polyolefin coated urea that in turn increases its rate of N release. Various investigators have studie d the effect of temperature on nutrient release rates of PCFs. Some studies reported that nutrient release rates of PCF doubled with a 10 C increase in temperature ( Kochba et al., 1990; Brown et al., 1966; Oertli and Lunt, 1962) In contrast, Huett and Gog el (2000) and Lamont et al. (1987 ) determined an increase in nutrient release rate of different PCFs of around 20 % and 15% respectively

PAGE 23

23 for a 10 C rise in temperature. Furthermore, Cabrera (1997) found a curvilinear relationship between the N release rat e of several PCFs and changes in average daily temperature up to a maximum temperature of 32C Nitrogen release was highly responsive to temperatures between 20 and 25C. Similarly, Husby et al. ( 2003 ) reported a dynamic N release response of PCFs to diur nal temperature changes from 20 to 40C. His results showed that increasing substrate temperature surrounding PCFs increases the rate of N release from the granules. There has been very little research conducted to describe the effect of soil physical and chemical properties on nutrient release rates of coated fertilizers. Nutrient release rates are not considered to vary directly with soil moisture content since moisture enters the coating mainly as water vapor rather than liquid. Kochba et al. ( 1990 ) repo rted no difference in nutrient release rate with moisture content at 50 to 100% of field capacity, slower release for soil at 25% of field capacity, and no release in dry soil. Although the mass movement of water and the diffusion of ions in the soil is pr oportional to soil moisture content the investigators attributed the discrepancy between such results and the expected effect on diffusion to the fact that lowering the soil moisture content within the range of field capacity has only a slight effect on t he vapor pressure of water in the soil. Similarly, Lunt and Oertli 1962 reported that moisture content exceeding the range of permanent wilting percentage to field capacity did not appreciably affect the N release rate of a coated fertilizer in a loam soil The investigators proposed that the diffusion rate depend s on the concentration gradient across the coating, and thus, they concluded that the differential ion concentration across the coating in this r ange of soil moisture was sufficient to maintain a c onstant N

PAGE 24

24 release rate. Friedman and Mualem (1994) used a model to quantitatively analyze the effect of soil moisture on nutrient release rate taking into account the release of nutrients from the coated fertilizer and their movement through th e soil by c hemical diffusion. The model showed that moisture content significantly affects nutrient release rates but that this effect is not linear. However, Christianson (1988) concluded that soil moisture content had a pronounced effect on the initial N release ra te of a PCF (7 d of incubation) but this effect was very minor once the granules became wet and N release started. Giordano and Mortvedt (1970) reported that N release from SCU decrease d with increasing soil water content. Rate of N release was much great er in moist than in flooded soil due to the formation of ferrous sulfide in saturated conditions that may seal the particle surface and greatly delay N release. Soil type has little influence on the nutrient release rate of coated fertilizers. Prasad and W oods (1971) compared N r elease rates of IBDU, UF hoof and horn meal and PCFs in virgin sphagnum moss peat and sand using a leaching technique during a 14 week period. They concluded that total N release d was higher in sand than in peat and this differen ce was more pronounced for IBDU, UF, and hoof and horn meal. Lower N released in p eat relative to sand w as attributed to the exchange capacity of peat for ammonium in the case of IBDU and to the lower microbial activity of the sterile peat that delay ed th e initial microbial decomposition of hoof and horn meal and UF, which slow ed down the nitrification of all fertilizers. Salman et al. ( 1989 ) studied the N release rate of a polymer coated urea (PCU) and SCU in a sandy and a wetland soil during 7 weeks. Soi l type did not influence N release rates from PCU. However, the total N released from SCU in the wetland soil doubled, which was likely due to faster

PAGE 25

25 breakdown of the S coating by microbes the activity of which was enhanced by favorable pH (~6.8) and high er organic matter content compared with the sandy soil. Similarly, Wang and Alva (1996) showed no difference in total N leached from PCU incubated in soils of similar particle size distribution. However, they found faster urea hydrolysis from the soil with higher solution pH and organic matter content. Although pH has no influence on total N release f rom coated fertilizers, pH has an effect on nitrification and thus, on the relative proportions of NH 4 N and NO 3 N (Oertli and Lunt, 1962) Methods to E stima te Nutrient Release Pattern s of EEF Estimating n utrient release patterns of EEFs is necessary to assess the effectiveness of these fertilizers to provid e nutrients that match crop nutrient demand. The slow release behavior of u rea aldehyde condensation pro ducts is affected by environmental factors such as biological activity, soil properties, soil moisture content, and temperature. Consequently, N release curves for these fertilizers vary according to these environmental conditions, resulting in a release p attern that differs from the sigmoidal form of nutrient uptake by plants. Little attention has been directed to model ing N release of these fertilizers because of the great variation in release performance (Shaviv, 200 1 ). In contrast, a variety of predicti on models have been developed to measure nutrient release rates of coated fertilizers because of their controlled release of nutrients and l ower sensitivity to soil conditions. The development of models to estimate N release from coated fertilizers is base d on three approaches: empirical ( Jarrel and Boersma, 1979; Kochba et al., 1990 ), mechanistic (Shaviv et al., 2003a; Shaviv et al., 2003b; Du et al., 2006; Du et al., 2008), and semi empirical (Gandeza et al., 1991; Zvomuya et al, 2003 ; Fujinuma et al., 20 09 )

PAGE 26

26 Various techniques conducted in field and laboratory settings have been used to verify nutrient release pattern of EEFs. Field methods allow expos ure of the fertilizer to fluctuating environmental conditions such as rainfall heat, and microbial activ it y in the soil. Common field techniques used to measure N release characteristics of EEFs include the closed or opened top cylinder system and the buried bag incubation Different versions of these methods have been implemented according to the objective s of the investigator For example, Sim m one and Hu tchinson (2005) and Huett (1997a ) incubated coated fertilizers in pots placed in the field and measure d the amount of N recovered after various leaching events. Hanafi et al. ( 2002 ) used an open field leach ing system where the coated fertilizers were incubated in lysimeters buried in the soil for 90 days. However, the most common method is to place a known amount of fertilizer in a mesh bag and bury it in the field (Gandeza et al., 1991; Zvomuya et al., 2003 ; Medina et al., 200 8 ; Wilson et al., 2009 ) Once the bags are retrieved from the field, the amount of N released is measured either by direct chemical analysis of the fertilizer granules or by the ir loss of weight. Laboratory methods to measur e nutrient release of EEFs are usually performed in the absence of plants and involve the incubation of the fertilizer under a specific set of conditions e.g. incubation media, temperature water content, sample preparation procedures and incubation period Hart e t al. ( 1994 ) classified laboratory soil incubation procedures as static methods (aerobic and anaerobic) and a erobic dynamic methods Savant et al. ( 1982 ) describe d a stagnant anaerobic incubation technique to measur e urea release from different coated fert ilizers using simulated physical, chemical, and microbiological conditions of wetland soils. An aerobic dynamic soil incubation

PAGE 27

27 procedure was initially proposed by Stanford and Smith (1972) to measure the N mineralization potential of soils. A series of la boratory soil incubation leaching technique s have been developed based on the Stanford and Smith method and modified accordingly to the needs of the stud y ( Alva, 1992; Paramasivam and Alva, 1997; Hanafi et al., 2002; Sartain et al., 2004; Broschat and Moor e, 2007; Fujinuma et al., 2009 ) These procedures consist o f the aerobic long term incubation of fertilizers at a predetermined temperature (usually ambient) with continuous or intermittent leaching of the nutrients released from the fertilizer at differen t time intervals for a specific period Various laboratory tests to characterize nutrient release of EEFs in water have also been developed over the years. These methods provide a more rapid technique to verif y nutrient release patterns of EEFs compared wi th lab soil incubation procedures. AOAC International (1965) developed the Activity Index (AI) to measure N release patterns of UF condensation products The AI measures the three different fractions in UF that differ in solubility in cold water (25C) and hot water (100C). N release patterns are described by Fraction I which represents the proportion of N slowly released and Fraction II which is the proportion of N that gradually releases during a period of 3 to 4 months, depending on the type of produ ct (Trenkel, 1997). The Tennessee Valley Authority (TVA) developed the 7 d dissolution rate (7 d DR) method to characterize N release from SCU. This static test involves immersing the fertilizer particles in water at 37.8C (100F) for 7 d. The total amoun t of N that dissolves in water after 7 d is the reported test value (McClellan and Sceib, 1973). The 7 d DR method has been extensively used to evaluate the release of urea from SCU granules (Blouin et al., 1971; Jarrell and Boersma 19 79 ). Several water b ased characterization procedures have

PAGE 28

28 literature reports a variety of static and dynamic incubation techniques with PCFs in water at various temperatures for different periods of t ime ( Al Zahrani 2000; Novillo et al., 2001; Tomaszewska and Jarosiewicz, 2002; Jarosiewicz an d Tomaszewska, 2003; Dai et al., 2008). Manufacturers generally used the time needed to release 75 to 80% of a given nutrient at temperatures ranging from 21 to 2 5C as an indicator of the duration of release (Shaviv, 2001). Similarly, Trenkel (2010) describe d laboratory methods used to test PCFs in China and Japan that basically involve the static dissolution of nutrients in water at temperatures ranging from 25 t o 100C. The objectives of this research were 1) to evaluate the effect of changes in sand/soil ratio, incubation temperature, and soil type on the effectiveness of the long term soil incubation column leaching technique developed by Sartain et al. (2004) to estimate N release rates of EEFs with time; 2) to assess the effect of extraction temperature, fertilizer sample size, and extraction time on the ability of an accelerated laboratory extraction method developed by Sartain et al. (2004) to estimate N, P and K release rates of EEFs; 3) to perform ruggedness testing on the performance of the accelerated lab extraction method to estimate N release patterns of EEFs; and 4) to develop a prediction model using data from the accelerated lab extraction method to predict the 180 d N release curve for a specific EEF product

PAGE 29

29 CHAPTER 2 EVALUATION OF A SOIL INCUBATION METHOD TO CHARACTERIZE NITR OGEN RELEASE PATTERNS OF ENHANCED EFFICIENCY FERTILIZE RS Enhanced efficiency fertilizers (EEF s ) are designed to gradually release nutrients in a pattern that closely matches nutrient demand by plants while potentially reducing nutrient losses to the environment through leaching, volatilization and/or runoff. Determin ing EEF nutrient release patterns is essential in the agrono mic evaluation of these materials. Understanding the rate of nutrient release from EEFs will aid in the selection of proper single or mixed EEF formulations for different environmental conditions. Nutrient release patterns of EEFs can be evaluated under fi eld or laboratory settings. When evaluated in the field, environmental factors that affect nutrient release such as soil temperature, moisture, pH, and microbial activity vary ; thus nutrient release measurements depend on current conditions. On the other h and, when evaluated in the laboratory, factors affecting nutrient release are controlled and set at specific optimal values, thus measurements are carried out under potential conditions Assessment of the N release profile of EEFs in the field is important for the design of effective soil management and fertilization practices. Field methods to estimat e net N release rates include the closed or opened top cylinder system and buried bag incubation. The major advantages of these techniques are : 1) they allow for the evaluation of the impact of on site environmental conditions on net N release rates and 2) they can also be used in both surface soils and subsoils with minimum disturbance of soil structure (Hart et al., 1994). Variations of these methods have b een evaluated in field studies to measure the N release rates of EEFs (Huett 1997a; Huett 1997b; Yanai et al., 1997; Hanafi et al., 2002; Zvomuya et al., 2003 and Chen et al., 2008).

PAGE 30

30 Despite the usefulness of in field techniques to investigate the actual agronomic effectiveness of EEFs, there is also a need to establish a lab method capable of rapidly and accurately evaluating the N release characteristics of EEFs. Laboratory methods to measur e net N transformation rates usually involve soil incubation for a defined period of time under specific conditions. These laboratory procedures include short term static incubation methods (aerobic and anaerobic) and long term aerobic dynamic incubation methods (Canali and Benedetti, 2006). These incubation technique s are usually simple, rapid, reproducible in large scale and easy to recreate in the laboratory under environmental conditions similar to those found in the field. Various laboratory techniques have been used to evaluate the N release profile of EEFs (Sav ant et al., 1982; Alva 1992; Paramasivam and Alva 1997; Broschat and Moore 2007). Previous work has focused mostly on develop ing soil incubation procedures ; however, there is a lack of information regarding the effectiveness of these methods to charact erize N release patterns of EEF s with respect to changes in incubation system environmental factors e g soil microbial population, moisture, temperature and soil type. The objective of this study was to e valuate the effect of changes in sand/soil ratio incubation temperature and soil type on the effectiveness of the Sartain et al. (2004) laboratory soil incubation method to estimate N release rates of EEFs with time The following hypotheses were tested in this study: 1) N release rates of the SRF s in crease with increasing amo unt of soil present in the incubation mixture and vice versa; 2) N release rates of the CRF s are not affected by the amount of soil present in the incubation mixture ; 3 ) N release rates of the EEFs increase with increasing temper atures; and 4 ) s oil texture does not influence N release rates of the EEFs.

PAGE 31

31 Materials and Methods A soil incubation column leaching technique was evaluated to determine its suitability as a standard method to estimate the long term N release patterns of EE F s The soil incubation technique developed by Sartain et al. (2004) was used as the standard method. This leaching column technique is based on the aerobic incubation of a sand/soil/fertilizer mixture exposed to intermittent leaching during a 180 d period These mixtures included: (1) 1710 g of non coated quartz sand ; (2) 90 g of the surface (0 to 15 cm) layer of a typical soil from central Florida such as Arredondo fine sand (Loamy siliceous, hyperthermic, Grossarenic Paleudult) ; and (3) the equivalent of 450 mg N from the EEF under evaluation. The components were mixed placed i n lysimeters and allowed to incubate for 180 d. The incubation lysimeters were constructed from cylindrical PVC pipe measuring 30 cm long by 7.5 cm diameter (Figure 2 1). M ixture s were retained in the lysimeters by glass wool mesh placed at the bottom. The sand/soil/ fertilizer mixture was initially wetted to 10% moisture content by adding 180 mL of water An ammonia trap consisting of 20 mL of 0.2 M H 2 SO 4 contained in a 50 mL beake r was placed in the head space of the incubation lysimeter. T h is solution was replaced and analyzed for NH 4 N by titration every 7 d to determine volatile N. The columns were incubated at about 21C in the laboratory. Nine leachings were collected a t 7, 14 28, 42, 56, 84, 112, 140 and 180 d from the start with one pore volume of 0.01% citric acid (500 mL). The use of citric acid aids in pH control (minimize s volatilization) and provides carbon to the microbes. T he columns were leached by gravity and excess solution was removed under vacuum for 2 min. The use of vacuum after every leaching event and the g aseous interchange through the open PVC spigot connected to the bottom cap w ere sufficient to maintain the system under aerobic

PAGE 32

32 conditions and thus ensure m icrobial activity and limit denitrification losses. At each leaching, t he l eachate volume was recorded and an aliquot was frozen for later analysis of total N by combustion using a n Antek 9000 N analyzer ( PAC Co., Houston, TX). In addition the pH of the l eachate and electrical conductivity (EC) w ere measured. Electrical conductivity was used to give an idea of the amount of total dissolved N in the leachate, therefore, EC data was not presented. The weight of total N recovered in the leachate was determine d by multiplying N concentration by leachate volume. At every sampling interval, the total N released with time was estimated by adding the total N present in the leachate and the volatile N Three factors (sand/soil ratio, incubation temperature and soi l type) that can potentially affect N release of the soil incubation column leaching method were selected to investigate the validity of this technique to measure N release pattern s of EEFs. In each study, only the variable under evaluation was modified wh ile all other parameters were maintained the same as the standard soil incubation method described above E ight EEFs described in Table 2 1 were selected to incorporate a wide range of commercially available EEF technologies. The fertilizers were separat ed in two groups: (1) controlled release fertilizers (CRF) that include sulfur coated urea (SCU), resin coated NPK, polymer sulfur coated urea (Poly S), reactive layer coated urea (RLCU) and polyolefin coated NPK; and (2) slow release fertilizers (SRF) tha t include IBDU, ureaform and biosolid s All the EEFs were used in the first and second studies while the SRF material s were omitted in s tudy 3. These studies were conducted from f all 2007 to f all 2009

PAGE 33

33 Study #1: Sand/Soil Ratio Effect The effect of sand/s oil ratio on the N release pattern of eight EEFs was determined by incubating sand/soil/fertilizer mixtures containing different quantities of soil. The standard method for incubation and leaching described previously was followed in this study ( Sartain et al. 2004) The total sand/soil mixture of 1800 g plus 450 mg N from each EEF were kept constant. Three experiments were conducted simultaneously to determine the sensitivity of the method to sand/soil ratio: 1) standard soil sample size (1710 g sand: 9 0 g soil); 2) low soil sample size (1755 g sand: 45 g soil) and 3) high soil sample size (1620 g sand: 180 g soil). The incubation medi um was composed of sand to allow for sufficient oxygen availability in the system and soil to serve as a biological inoc ulant. The purpose of adding native field soil in the mixture was to introduce the biotic component of a natural system. The nutrient release mechanisms of EEFs depend on microbial activity temperature and soil moisture content. The quantity of inoculant soil used in the incubation media could influence the water content and microbial activity within the system, thus affect ing N release rate. The volumetric water content of each sand/soil mixture was measured using a HydroSense Soil Water Content Measure ment System (Campbell Scientific, Inc., Logan, UT) with a 20 cm rod. The measurements were made 3 d after the initiation of the experiment and at every leaching event to allow for stabilization of the system, and they continued until the third leachate (28 d). Study #2: Incubation Temperature Effect The influence of temperature on the rate of N release was studied by incubating the sand/soil/fertilizer mixtures at temperatures ranging from 21 to 35C. Three experiments to determine the N release rate o f ei ght EEFs were conducted

PAGE 34

34 simultaneously in separate growth chambers. Incubation lysimeters were prepared as described in the standard soil incubation column leaching technique. Lysimeters were placed in growth chambers at the following incubation temperatur es: 1) standard temperature (21C); 2) medium temperature (25C) and 3) high temperature (35C). After incubati ng for a specified time the columns were removed from the growth chambers and leached following the standard leaching procedure in order to esti mate the amount of N released. Study #3: Soil Type Effect The effect of soil types other than Arredondo fine sand on the transfer of N to the soil was investigated using the soil incubation leaching method. The soil columns were prepared as described in th e standard method and incubated at about 21C in the laboratory for 180 d. Since the intent of this study was to develop a generic soil incubation method that could be used anywhere, it was important to determine whether soil type influences the method erformance. Soils were obtained from four locations representing major geographic regions of the United States. Four experiments were performed to evaluate the N release patterns of six EEFs incubated in soils from: 1) Florida; 2) California; 3) Pennsylvan ia and 4) Iowa. At each site, surface (top 15 cm) soil was collected, air dried and passed through a 2 mm sieve. T he chemical and physical properties of the soils were characterized using triplicate samples (Table 2 2). Total carbon (C) and total N were d etermined at the Wetland Biogeochemistry Laboratory at the University of Florida using a Flash EA 1121 NC soil analyzer (Thermo Electron Corporation, Waltham, MA). Particle size distribution of the soils was obtained using the method described in the USDA Soil Survey Laboratory Methods Manual (2004).

PAGE 35

35 S tatistical Analysis All studies included four non amended columns and four columns containing water soluble ammonium nitrate (AN) as a control treatment. Treatments were replicated four times in a randomized c omplete block design. Statistical analysis of data was performed separately for each study using Statistical Analysis System (SAS) software (version 9.0; SAS Institute, Cary, NC). The pH data from the fertilizer treatments were statistically analyzed using an analysis of variance procedure. Nonlinear regression curves were fitted to the N release data separately for each EEF material to develop N release curves. The Gaussian Newton method of SAS PROC NLIN was used to calculate the non linear regressions. Th e non linear regression curves follow ed the equation where N(t)= cumulative percent age N released (asymptote of the curve); b = intercept and k = rate, with units of percent age N release d per d. For each study, the means of the estimated coefficients of the N release curves were compared with the general linear model procedure (PROC Results and Discussion Study #1: Sand/Soil Ratio Effect The percentages of N released after 180 d of incubation for all EEF materials are shown in Figures 2 2 and 2 3. Despite the differences in N release rates among the three sand/soil ratio s similar N release patterns with time were obtained for each fertilizer material. Regardless of the sand/soil ratio, 93% of the N from the water soluble AN was released after 7 d of incubation and 96% was recovered by the time of the second leach ing (14 d). The pH of the AN leachate did not change substantially during the study. It remained between 6.5 and 7.0 depending on the sand/soil ratio.

PAGE 36

36 Columns were leached with a 0.01% citric acid solution that had an average pH of 3.4 T herefore the pH of the leachate should be to some extent lower than would be expected from the soil solution leached from these columns. C itric acid is a weak acid that partially io nizes in solution, so it only releases a portion of its hydrogen ions when added to the sand/soil/fertilizer mixture. In addition citric acid was expected to have a short lifetime in the mixture since it can be used by microbes as a carbon source. Therefo re, the addition of the 0.01% citric acid solution was not expected to cause a pronounced, long term effect to decrease the pH of the mixture. This mild effect of the citric acid on the mixture pH was further confirmed by the reduction of less than one uni t in the leachate pH of the non amended soil (control) columns until the end of the experiment. The leachate pH from the control columns varied from 6.3 to 7.2 6.5 to 7.2 and 7.0 to 7.6 for the low, standard and high soil sample size s, respectively. Ac ross the three sand/soil ratio s all EEFs followed basically the same trend showing a reduction of pH with time (Figure 2 4 ). Initially the pH of the leachates from all EEFs was about 7.0 or above, but the values decreased gradually with time reaching a l eachate pH of below 4.0 for ureaform and all CRFs and above 6.0 for IBDU and the biosolid s at the end of the experiment. Ureaform behaved differently when 45 g of soil were used in the mixture, reaching a final pH of approximately 6.5. The initial high pH of about 8.0 from soil columns amended with urea based EEFs was likely due to hydrolysis of urea into ammonium carbonate through the action of the urease enzyme. The leachate pH then decreased as a result of production of nitrate through nitrification (Pa ramavisan and Alva, 1997). N itrogen was recovered from the soil columns amended

PAGE 37

37 with CRFs despite a pH below 4.0 observed during the last leaching events since their nutrient release mechanism is not pH dependent Volatile ammonia N was captured from SCU, ureaform and the biosolid s material for all sand/soil ratios. For the duration of the experiments, ammonia N was detected only during the first 7 d of incubation for the standard and low soil sample size experiments and during 28 d for the high soil samp le size experiment. Regardless of the sand/soil ratio, ammonia volatilization average d 2.2 and 2.7% of the total N released from SCU and biosolid s, respectively. The rate of ammonia volatilization from ureaform was influenced by the quantity of soil in the incubation mixture. During the 180 d incubation period, ammonia N represented roughly 5.4, 7.4 and 7.7% of the total N released from ureaform amended columns receiving 45, 90 and 180 g of soil respectively. In addition, a small amount of ammonia N (1.6 % of total N released) was volatilized from IBDU when incubated in columns containing 180 g of soil. Regression coefficients, R 2 and P values of the non linear regression equations are provided in Tables 2 3 and 2 4 for the CRFs and SRFs, respectively. Tw o different N release rate patterns were observed between the CRFs and SRFs for all three sand/soil ratio experiments. When comparing the standard soil sample size with the low soil sample size treatment, no difference in the coef f icients a and (a b) of th e estimated N release curves was found in most CRFs except for SCU and RLCU. The SRFs showed a similar trend with no difference for IBDU and ureaform but a significant difference for the biosolid s material. On the other hand, when comparing the standard s oil sample size with the high sample size, there was generally no difference in the a and k coeficients of the estimated N release curve s but the (a b) coefficient was

PAGE 38

38 different for all CRFs. An opposite trend was observed with the SRFs which resulted in mostly highly significant (P<0.0001) difference s in all the coefficients of the estimated N release curves. There was a distin c t effect of the quantity of soil present in the incubation media on the rate of N release (Table 2 5 and 2 6) Overall, the init ial N release rates from columns innoculated with 45 g of soil were similar to those from columns receiving the standard soil samp l e size (90 g soil), while the total cumulative N released was either lower or similar depending on the fertilizer. In contra s t, the initial N release rates and the total cumulative N released from columns rece i ving 180 g of soil were higher compared with those receiving the standard soil sample size. This trend was followed mostly with the SRFs while CRFs resulted in hig h er ini tial N relase rates but similar total cumulative N released at the end of the experiment. At the beginning of the experiment, the volumetric water content of the systems containing 45 and 90 g of soil was the same at about 5% but was 1 % higher for the co lumns receiving 180 g of soil. After every leachate, the volumetric water content of the columns amended with 45 g remained constant, while it increased by 1% and 2% for the columns receiving 90 and 180 g of soil respectively. The difference in N released from columns receiving 45 and 180 g of soil could be explained partly by the variation in the moisture content of the incubation media, since more soil allow ed greater water retention while less soil likely caused dr ier conditions in the incubation system The resin coated NPK material had higher initial N release rates after 7 d of incubation when using 45 g of soil compared with the standard treatment. This high initial level of N may have be en due broken fertilizer particles, microscopic cracks or

PAGE 39

39 imper fections in the fertilizer coating and greater leaching of ammonium N since less soil was available to ret ain ammonium on the cation exchange sites. Similar trends were observed by Patel et al. ( 1 977 ) T here was no difference in N release rate for Poly S b etween the standard and low soil sample size s However, when comparing the standard with the high soil sample size treament, N release patterns for resin coated NPK and Poly S were different, with faster N release rates through 180 d from the columns recei ving 180 g of soil High er N release rates were seen after 7 d of incubation until the second leachate (14 d) This result sug gested that N release was influenced by soil moisture content mainly at the beginning of the incubation When the dry granules of these CR Fs we re placed in the moist incubation media, a high osmotic gradient develop ed between the fertilizer and the media causing water vapor to move through the coating, dissolve the N core and diffuse the N out of the fertilizer particle. Because of t he higher moi s ture content of the columns receiving 180 g of soil, the difusion of the ions away from the fertilizer granules was likely faster resulting in a very low N concentration surrounding the granule s This higher concentration gradient across the membrane consequently result ed in greater diffusion of N th r ough the coating. Th ese higher N release rates were expected to happen when N inside the granule was being dissolved at the beginning of the incubation continuing until all the N was dissolved w ithin the granule. Once the N concentration within the granule was reduced, the rate of diffusion was expected to decrease. Christianson (1988) also found that N re lease rates from a polymer coated urea incubated at different soil moisture levels varied du ring the fir s t 7 d of incubation, with N release increasing slightly as soil

PAGE 40

40 moisture content increased. Christianson concluded that once the fertilizer granules were moistened and N release started, soil moisture had little effect on N release rate. The s and/soil ratio of the mixture influenced the rate of N release from IBDU which increas ed with the quantity of soil present in the system. Environmental factors such as moisture, temperature and pH have been reported to greatly affect the hydrolysis and d ecomposition of urea aldehyde condensation products (Jahns et al., 2003). Therefore, a faster N release rate from the columns amended with 180 g of soil was expected because of the higher moisture content of th is system. Microbial activity of each incubati on mixture was quantified by measuring m icrobial respiration Microbial respiration was determined using a simple titrimetric method for measuring carbon dioxide (CO 2 ) evolution from the incubation mixtures after CO 2 entraptment in alkali (Anderson 1982) Although microbial respiration among the three incubation mixtures was not statistically different, the respiration rate from the system containing the low soil sample size (45 g soil) decreased by 5% while it increased 20% in the system having the high so il sample size (180 g soil) compared with the standard mixture (90 g soil). This result suggested that microbial activity was enhanced by adding more soil to the sand/soil/fertilizer mixture but not to a significant degree Similar patterns of N release w ere shown by ureaform and the biosolid s material, with N release rates increasing as the quantity of soil present in the incubation system increased. In addition, there was about a 50% increase in the initial N released between the standard and high soil s ample size treatments. Higher N release rates from columns receiving the high soil sample size treatment (180 g soil) could be explained by greater microbial activity that directly influence d the rate of decomposition of these fertilizers

PAGE 41

41 Likewise, the hi gher moisture content of th is mixture directly affected the hydrolysis rate of the biosolids material and indirectly promote d N mineralization in the media Although ureaform and the biosolids material released N at a steady rate, no more than 20% of the t otal N applied was recover ed after the initial N release regardless the sand/soil ratio This low recovery of N could be attributed to losses through denitrification during the first weeks of incubation ( Koivunen and Horwath 2004) or fixation of ammonium N on the exchange sites Study #2: Incubation Temperature Effect Approximately 97% of the N applied as water soluble AN was released at the 7 d leaching event T he release rate was not affected by temperature. L eachate pH obtained from the control columns oscillated between 6.9 and 7.2 In general, the initial pH of the leachates from all EEFs varied between 6.3 and 8.6 There was a continuous pH reduction with time until a pH 4.0 was reached for all CRFs and ureaform, and about 6.5 for IBDU and the biosol id s at 180 d (Figure 2 5) This pH reduction with time was likely due to the addition of the 0.01% citric acid solution (pH=3.5) and the process of nitrification. Volatile ammonia N was observed after 7 d of incubation only for ureaform, the biosolid s mate rial and SCU at all three temperature s I ncubation temperature had an influence on the rate of ammonia volatilization. Ammonia volatilization was similar at 21 and 25C in 180 d accounting for 7.9, 3.0 and 1.7% of the total N released from ureaform, bio solid s, and SCU respectively. However, volatilization rate increased in magnitude approximately 0.5, 3 and 10 times for ureaform, biosolid s, and SCU respectively when the columns were incubated at 35C. Volatile ammonia N was also detected during the f irst 14 d of incubation from IBDU and RLCU incubated at 25 and

PAGE 42

42 35C. Depending on temperature, ammonia volatilization ranged from 1.4 to 2.0% and 1.1 to 3.6% of the total N released from IBDU and RLCU respectively. The N release rate from EEFs was directl y related to incubation temperature. The regression coefficients for the fitted N release curves by temperature for all the CRFs and SRFs are shown in T ables 2 7 and 2 8, respectively. Compared with 21 C t he a and (a b) parameters of the fitted N release curves were generally not different at 25C while only (a b) was significantly different at 35C for most CRFs These results indicated that similar total cumulative N release d w as obtained at the end of the experiment regardless of temperature while the initial N release rates increased as the temperature increase d from 21C to 35C. In contrast, for p olyolefin c oated NPK both parameters [ a and (a b) ] of the fitted N release curve were different at 25 C and 35C compared with 2 1C Although temperature markedly affected the release of N from CRFs release rates did not increase proportionally with temperature (Figure 2 6 and 2 7) When comparing N release rates at 21 C with those at 25C, the release rate (k) parameter of the fitted N release curves almo st doubled for resin coated NPK, Poly S and RLCU (Table2 7) Similar variations in N release patterns from 20 to 25C were found by Cabrera (1997) in an N leaching study with similar coated fertilizers. Overall, t he effect of temperature at 35C compared with 21 C was reflected by higher initial N release rates that gradually declined after 28 d of incubation (Table 2 9) The increase in initial N released was 28% for SCU, but i t almost doubled for resin coated NPK and Poly S and tripled for RLCU. Th e s e r esults were similar to the doubling in the initial N release rate of polymer coated fertilizers reported by Oertli and Lunt (1962) for a 10 C increase in temperature

PAGE 43

43 Likewise, Huett and Gogel (2000) found a similar temperature effect on coated fertilizers incubated at 30 and 40 C where N was released unevenly throughout the incubation period with the highest rate occurring during the early part of the release period Brown et al. ( 1966 ) also showed a similar effect of temperature between 25 and 35C on re lease of urea from resin coated granules mixed with quartz sand. The behavior of polyolefin coated NPK was different compared with the other CRFs. Increasing the incubation temperature from 21 to 35C quintupled the initial N release rate, doubled the rele ase rate (k) and increased the total cumulative N released by 23% (Tables 2 7 and 2 9) This marked effect of temperature on N release rates was likely due to the strong dependence of the diffusion coefficient and solubility of N on temperature (Gandeza e t al., 1991). The effect of te mperature on the N release pattern of IBDU was mostly observed during the first 28 d of incubation (Table 2 10) At 35C, a more pronounced effect with doubling of the initial N release rate was ob served This result was confi rmed by a higher leachate pH of 8.4 at 35 C compared with pH at the lower temperatures ( ~ 7.1) after 7 d of incubation due to the faster dissolution and hydrolysis of IBDU. Furthermore, at 21 and 25 C, leachate pH increased by one unit at the second and thi rd leaching event indicating the gradual hydrolysis of IBDU with time (Figure 2 5) Lunt and Clark (1969) also demostrated a moderate enhancement of solubilization of IBDU by a temperature increase from 50 to 80F (10 to 27 C) N release rates from u reafo rm were similar at 21 and 25C On average, only 33% of the total N applied was recovered during the 6 months of the experiment from the columns incubated at 21 and 25C. The temperature effect was significant at 35C

PAGE 44

44 showing the highest release rate (20% of the N applied ) 7 d after incubation followed by a gradual decrease in the rate of N mineralization with time, reaching a total N released of 50% of the N applied (Table 2 10) The biosolid s material showed a slow and steady mineralization pattern at al l temperatures. A fter a high initial N release rate, only about 20 % of the N applied was mineralized until the end of the experiment regardless of temperature A n increase in initial N release (30%) and total cumulative N released (15%) was observed with b iosolid s at 35C compared with the lower temperatures (Table 2 10) The percent age N release d per day (k) from ureaform and biosolid s almost doubled by a 10 increase in temperature from 25 to 35C which led to the conclu sion that temperature positively a ffected microbial activity (Table 2 8) Study #3: Soil Type Effect O n average, 96% of the N applied as water soluble AN was released after 28 d of incubation from columns containing California, Pennsylvania and Iowa soils while 95% was released by the 14 d leaching event with the Florida soil. The leachate pH from the non amended columns ranged from 6.2 to 7.2. Initially, leachate pH oscillated from 6.2 to 8.4 It then gradually decreased with time until the termination of the experiment ranging from 3.4 to 5.0 depending on the soil type and fertilizer material (Figure 2 8) Volatile ammonia N was only detected from columns amended with SCU after 7 d of incubation. Ammonia volatilization varied from 1.1 to 2.2% of the total N released from SCU depending o n soil type. Compared with the Florida soil used in the standard method, s oil type influenced the percent age N release d per day ( k ) f or all EEFs (Table 2 11 ). In general, initial N release rates [ (a b) ] and total cumulative N released ( a ) were not signific antly different for any of the soils, but the release rate constant ( k ) was significantly different for all

PAGE 45

45 soils. The k value increased depending on soil type following the order: Iowa>California>Pennsylvania>Florida. The N release from EEFs incubated in the loamy Iowa and California soils procee ded more rapidly than it did in the silt loam and sandy soils ( Table 2 12 ). This result was likely due to the higher clay content of the Iowa and California soils which consequently provide d higher water holding capacity in the incubation system. Wang and Alva (1996) concluded that N release rates of polymer coated urea were similar among soils of similar particle size distribution H owever, they also suggested that urea hydrolysis and nitrification were faster in soils that had a greater amount of organic C Considering the loamy soils used in this study the Iowa soil had a greater C content (2.65%) than the California soil (2.09%) C onsequently it was expected that higher N release rates would be obtained from the columns incubated with the Iowa soil (Figure 2 9) Despite differences in the N release rate among soils, more than 70% of the N applied was released from all EEFs by the end of the 180 d incubation period regardless of soil type (Table 2 12) Fig ure 2 1. Incubation lysimeters.

PAGE 46

46 F igure 2 2 Nitrogen release curves f or SCU, resin coated NPK, P oly S and RLCU as affected by soil sample size using a soil incubation method

PAGE 47

47 Figur e 2 3. Nitrogen release curves f or polyolefin coated NPK, IBDU, ureaf orm and biosolid s as affected by soil sample size using a soil incubation method

PAGE 48

48 Figure 2 4 Effect of N source on leachate pH using the soil incubation method A= low soil sample size (1755 g sand: 45 g soil) ; B=standard soil sample size ( 1 710 g sand: 90 g soil) ; C=high soil sample size (1 620 g sand: 180 g soil)

PAGE 49

49 Figure 2 5 Effect of N source on leachate pH using the soil incubation method A= standard temperature ( 21 C) ; B=medium temperature ( 25C) ; C=high temperature ( 3 5C)

PAGE 50

50 F igure 2 6 Nitrog en release curves f or SCU, resin coated NPK, P oly S and RLCU as affected by incubation temperature using a soil incubation method

PAGE 51

51 Figure 2 7 Nitrogen release curves f or polyolefin coated NPK, IBDU, ureaform and biosolid s as affected by incubation t emperature using a soil incubation method

PAGE 52

52 Figure 2 8. Effect of N source on leachate pH u sing the soil incubation method. A=Florida soil; B=California soil; C=Pennsylvania soil and D=Iowa soil.

PAGE 53

53 Figure 2 9 Nitrog en release cu rves f or SCU, resin coated NPK, P oly S Polyolefin coated NPK and RLCU as affected by soil type using a soil incubation method

PAGE 54

54 Table 2 1. Enhanced efficiency fertilizer specifications. N Source Formulation (N P 2 O 5 K 2 O) Release duration (mont hs) 1 Principle source 2 N P 2 O 5 K 2 O Sulfur Coated Urea 39 0 0 1 2 WSON ----Resin Coated NPK 19 6 12 3 4 AN, AP AP,CP KS Polymer Sulfur Coated Urea 37 0 0 6 PSCU ------Reactive Layer Coated Urea 43 0 0 2 3 PCU ------Polyolefin Coated NPK 18 6 18 6 AN, AP KN AP,CP KN IBDU 31 0 0 2 3 WSON ------Ureaform 38 0 0 6 9 70% WIN 18%SAWSN 12% WSON ------Biosolid s 6 3 2 variable A B A B A B 1 Approximate at 21C soil temperature, as specified by the manufacturer. 2 WSON= water soluble organic N (urea); AN= ammonium nitrate;AP=ammonium phosphate; CP=calcium phosphate; KS= potassium sulfate; PSCU=polymer sulfur coated urea; PCU=polymer coated urea; KN=potassium nitrate WIN=water insolub le N ; SAWSN=slowly available water soluble N ; A B =activated biosolid Table 2 2. Physical and chemical characteristics of the soils. Soil pH Particle Size Total Origin Texture Clay Sand Silt N C (1:2) -------------------------% -----------------------Florida Sand 6.7 2.4 94.5 3.1 0.10 1.49 California Loam 7.1 19.7 48.8 31.5 0.10 2.09 Pennsylvania Silt loam 5.6 7.9 24.9 67.2 0.24 3.66 Iowa Loam 6.0 16.7 48.0 35.3 0.16 2.65

PAGE 55

55 Table 2 3. Regression analysis of estimated N release rate from diffe rent CRFs using the soil incubation method with three sand/soil ratios. Sand/soil Ratio 1 a (a b) k R 2 P value ------------Sulfur Coated Urea -------------1) Standard 78.191 48.597 0.0184 0.99 <0.0001 (1.185) 2 (1.241) (0.0017) 2) Low 69.850 39.51 6 0.0313 0.98 <0.0001 (4.394) (4.190) (0.0024) 3) High 74.092 39.650 0.0208 0.99 <0.0001 (4.441) (3.439) (0.0009) 1 vs 2 ** *** 1 vs 3 NS ** NS -----------Resin Coated NPK -----------1) Standard 97.041 83.775 0.0223 0.99 <0.0001 (7. 918) (7.523) (0.0014) 2) Low 95.464 69.727 0.0337 0.99 <0.0001 (1.951) (2.131) (0.0019) 3) High 104.771 65.305 0.0245 0.99 <0.0001 (3.824) (7.252) (0.0019) 1 vs 2 NS *** 1 vs. 3 NS NS -----------Polymer Sulfur Coated Urea ----------1) Standard 83.435 74.051 0.0105 0.99 <0.0001 (0.528) (1.112) (0.0005) 2) Low 74.465 65.823 0.0147 0.99 <0.0001 (8.264) (7.849) (0.0030) 3) High 85.034 73.504 0.0170 0.99 <0.0001 (3.699) (4.107) (0.0046) 1 vs. 2 NS NS NS 1 vs. 3 NS N S NS -----------Reactive Layer Coated Urea -----------1) Standard 88.248 91.134 0.0306 0.99 <0.0001 (4.167) (2.690) (0.0006) 2) Low 80.599 81.739 0.0310 0.99 <0.0001 (2.505) (2.701) (0.0010) 3) High 88.382 81.967 0.0267 0.99 <0.0001 (6. 888) (8.544) (0.0012) 1 vs. 2 NS 1 vs. 3 NS NS

PAGE 56

56 Table 2 3. Continued. Sand/soil Ratio 1 a (a b) k R 2 P value -----------Polyolefin Coated NPK -----------1) Standard 96.136 101.153 0.0079 0.99 <0.0001 (1.932) (2.311) (0.0009) 2) Low 96.691 104.196 0.0145 0.99 <0.0001 (1.756) (1.752) (0.0009) 3) High 108.081 111.103 0.0071 0.99 <0.0001 (2.218) (4.057) (0.0004) 1 vs. 2 NS NS ** 1 vs. 3 *** ** NS 1 1= 1710 g sand: 90 g soil; 2=1755 g sand: 45 g soil and 3=1620 g sa nd: 180 g soil. 2 Standard Deviation is shown in parenthesis. 3 where a, (a b)= regression coefficients and k = release rate. 4 NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.001.

PAGE 57

57 Table 2 4. Regression analysis of estimated N release rate from different SRFs using the soil incubation method with three sand/soil ratios. Sand/soil Ratio 1 a (a b) k R 2 P value ------------IBDU -------------1) Standard 77. 125 82.631 0.0415 0.98 <0.0001 (2.707) 2 (3.396) (0.0022) 2) Low 73.526 76.343 0.0346 0.98 <0.0001 (1.065) (5.174) (0.0031) 3) High 102.388 120.695 0.0403 0.99 <0.0001 (2.851) (3.484) (0.0012) 1 vs 2 NS NS ** 1 vs 3 *** *** NS ----------Ureaform -----------1) Standard 36.904 22.844 0.0078 0.98 <0.0001 (0.489) (1.434) (0.0014) 2) Low 32.639 19.910 0.0067 0.99 <0.0001 (1.228) (1.997) (0.0007) 3) High 78.743 57.077 0.0022 0.99 <0.0001 (10.929) (10.185) (0.0005) 1 vs 2 NS NS NS 1 vs 3 *** *** *** -----------Biosolid s -----------1) Standard 59.140 19.280 0.0196 0.93 0.0004 (1.486) (1.324) (0.0060) 2) Low 54.305 12.139 0.0119 0.89 0.0013 (1.802) (1.482) (0.0059) 3) High 74.040 14.437 0.0150 0.99 <0. 0001 (4.356) (1.465) (0.0045) 1 vs 2 *** NS 1 vs 3 *** ** NS 1 1= 1710 g sand: 90 g soil; 2=1755 g sand: 45 g soil and 3=1620 g sand: 180 g soil. 2 Standard Deviation of the coefficients is shown in parenthesis. 3 whe re a, (a b)= regression coefficients and k = release rate. 4 NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.001.

PAGE 58

58 Table 2 5. Effect of sand/soil ratio on total N released from five CRFs with time Total N Releas ed (% applied) Time (days) Cum. Sand/soil Ratio 1 7 14 28 42 56 81 112 140 180 (%) ------------Sulfur Coated Urea -------------1) Standard 33.9 8.0 8.3 5.3 4.9 5.7 6.4 2.2 1.9 76.5 2) Low 37.4 8.0 7.2 7.4 3.2 2.8 1.3 2.2 1.5 71.1 3) High 39.9 3.1 10.1 5.6 3.9 2.9 2.5 4.0 3.0 75.0 1 vs. 2 NS NS NS *** *** *** NS NS NS 1 vs. 3 ** NS NS NS *** *** ** NS -----------Resin Coated NPK -----------1)Standard 26.2 9.8 13.7 15.7 7.3 12.8 5.2 2.4 1.6 94.7 2) Low 39.7 13.0 15.8 10. 1 5.8 6.3 2.8 1.7 0.7 95.9 3) High 51.7 4.5 15.5 9.7 6.9 9.9 2.4 1.7 0.9 103.2 1 vs. 2 ** NS ** ** *** ** ** ** NS 1 vs. 3 *** ** NS ** NS ** ** ** NS -----------Polymer Sulfur Coated Urea -----------1)Standard 12.7 6.8 9.2 11.2 3.0 7.2 9.7 4.0 11.3 75.1 2) Low 13.5 7.9 10.4 7.7 6.5 6.6 7.3 5.4 5.3 70.5 3) High 19.0 7.9 13.8 8.6 6.8 10.8 4.0 4.8 7.0 82.9 1 vs. 2 NS NS NS ** *** NS ** NS *** NS 1 vs. 3 *** NS *** ** *** NS ** NS -----------Reactive Layer Coated Urea -----------1)Standard 14.6 14.4 19.5 16.9 7.3 5.1 7.9 1.7 1.9 89.1 2) Low 16.1 9.0 21.9 12.0 7.8 7.0 2.7 1.5 4.6 82.7 3) High 24.1 3.5 20.3 14.9 9.0 9.5 2.5 1.8 1.4 87.0 1 vs. 2 NS *** ** NS ** *** NS *** NS 1 vs. 3 *** *** NS NS *** *** NS NS NS -----------Polyolefin Co ated NPK -----------1)Standard 2.8 1.7 7.8 10.1 9.8 12.2 10.6 6.2 9.8 71.1 2) Low 2.5 2.2 15.3 11.7 9.5 12.5 8.9 7.2 7.1 76.9 3) High 4.8 2.1 9.2 7.7 9.6 15.8 10.5 8.5 8.2 76.3 1 vs. 2 NS NS *** NS NS NS ** ** *** 1 vs. 3 *** NS NS NS ** NS *** *** 1 1= 1710 g sand: 90 g soil; 2=1755 g sand: 45 g soil and 3=1620 g sand: 180 g soil. 2 Single degree of freedom contrast s were generated using SAS GLM Proc. NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0 .0001.

PAGE 59

59 Table 2 6. Effect of sand/soil ratio on total N released from three SRFs with time Total N Released (% applied) Time (days) Cum. Sand/soil ratio 1 7 14 28 42 56 81 112 140 180 (%) ------------IBDU -------------1)Standard 13.9 18.7 20.5 8.7 6 .1 3.5 5.0 1.6 1.2 79.2 2) Low 17.9 2.8 22.1 15.1 7.4 5.3 0.9 0.6 0.0 72.0 3) High 14.0 16.2 30.2 23.8 7.6 5.2 4.4 0.0 0.0 101.4 1 vs. 2 *** NS ** NS NS *** ** *** 1 vs. 3 NS *** ** *** NS NS NS *** *** *** -----------Ureaform -----------1) Standard 14.6 1.1 3.9 1.4 1.4 1.4 2.7 1.6 3.7 31.9 2) Low 13.3 1.3 1.6 1.5 1.4 1.4 2.7 1.7 1.7 26.5 3) High 21.9 2.0 1.6 1.0 0.8 3.2 2.5 2.6 4.1 39.8 1 vs. 2 NS NS *** NS NS NS NS NS *** *** 1 vs. 3 *** ** *** ** *** NS ** *** -----------Biosoli d s -----------1)Standard 39.7 7.4 2.2 1.4 1.3 1.4 2.8 1.3 2.1 59.6 2) Low 41.1 4.4 1.4 0.0 0.0 1.7 1.4 1.5 1.1 52.6 3) High 59.3 4.5 2.1 0.8 0.6 1.3 2.9 0.3 1.8 73.6 1 vs. 2 NS *** *** *** ** 1 vs. 3 *** NS ** *** NS NS *** NS ** 1 1= 17 10 g sand: 90 g soil; 2=1755 g sand: 45 g soil and 3=1620 g sand: 180 g soil. 2 Single degree of freedom contrast s were generated using SAS GLM Proc. NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.0001.

PAGE 60

60 Table 2 7 Regression analysis of estimated N release rate from different CRFs using the soil incubation method at three incubation temperatures. Temperature 1 a (a b) k R 2 P value ------------Sulfur Coated Urea -------------1) 21C 74.286 45.988 0.0224 0.99 <0.0001 (0.833) 2 (1.293) (0.0018) 2) 25C 68.073 44.194 0.0299 0.97 <0.0001 (4.278) (5.501) (0.0052) 3) 35C 79.608 36.923 0.0103 0.99 <0.0001 (2.846) (2.049) (0.0019) 1 vs 2 NS NS 1 vs 3 NS ** ** -----------Re sin Coated NPK -----------1) 21C 97.041 83.775 0.0223 0.99 <0.0001 (7.918) (7.528) (0.0014) 2) 25C 94.394 73.884 0.0366 0.99 <0.0001 (1.586) (3.305) (0.0022) 3) 35C 101.330 69.890 0.0431 0.99 <0.0001 (2.234) (4.967) (0.0028) 1 vs 2 NS *** 1 vs 3 NS ** *** -----------Polymer Sulfur Coated Urea ---------1) 21C 83.435 74.051 0.0105 0.99 <0.0001 (0.528) (1.112) (0.0005) 2) 25C 73.747 67.648 0.0283 0.99 <0.0001 (3.090) (2.923) (0.0029) 3) 35C 79.340 66 .075 0.0307 0.99 <0.0001 (7.324) (6.885) (0.0052) 1 vs 2 NS *** 1 vs 3 NS ***

PAGE 61

61 Table 2 7 Continued. Temperature 1 a (a b) K R 2 P value -----------Reactive Layer Coated Urea -----------1) 21C 88.248 91.134 0.0306 0.99 <0.0001 (4.167) (2.690) (0.0006) 2) 25C 80.681 72.081 0.0530 0.99 <0.0001 (3.718) (3.026) (0.0019) 3) 35C 88.893 74.370 0.0569 0.99 <0.0001 (11.623) (5.663) (0.0042) 1 vs 2 NS *** *** 1 vs 3 NS ** *** -----------Polyolefin Coated NPK ----------1) 21C 96.136 101.153 0.0079 0.99 <0.0001 (1.932) (2.311) (0.0009) 2) 25C 86.535 96.519 0.0145 0.99 <0.0001 (2.144) (2.786) (0.0004) 3) 35C 89.320 82.745 0.0175 0.99 <0.0001 (3.323) (5.958) (0.0027) 1 vs 2 ** ** 1 vs 3 ** ** *** 1 21C =standard ; 25C =medium and 35C =high 2 Standard Deviation of the coefficients is shown in parenthesis. 3 where a, (a b)= regression coefficients and k = release rate. 4 NS = not significant, *= significan t P<0.05, **= significant P<0.01 and ***= significant P<0.001.

PAGE 62

62 Table 2 8 Regression analysis of estimated N release rate from different SRFs using the soil incubation method at three incubation temperatures. Temperature 1 a (a b) k R 2 P val ue ------------IBDU -------------1) 21C 77.125 82.631 0.0415 0.98 <0.0001 (2.707) 2 (3.396) (0.0023) 2) 25C 78.546 99.827 0.0628 0.99 <0.0001 (2.500) (9.303) (0.0080) 3) 35C 77.579 68.352 0.0606 0.99 <0.0001 (7.153) (4.588) (0.0051) 1 vs 2 NS ** ** 1 vs 3 NS ** ** -----------Ureaform -----------1) 21C 36.904 22.844 0.0078 0.98 <0.0001 (0.489) (1.434) (0.0014) 2) 25C 53.766 36.953 0.0034 0.99 <0.0001 (8.187) (4.577) (0.0006) 3) 35C 65.786 46.557 0.0061 0.99 <0.000 1 (2.518) (1.139) (0.0002) 1 vs 2 ** *** *** 1 vs 3 *** *** -----------Biosolid s -----------1) 21C 59.140 19.280 0.0196 0.93 0.0004 (1.486) (1.324) (0.0060) 2) 25C 57.057 18.517 0.0192 0.95 <0.0001 (7.523) (1.079) (0.0019) 3) 35C 66.519 17.027 0.0331 0.90 0.0011 (7.718) (0.702) (0.0064) 1 vs 2 NS NS NS 1 vs 3 1 21C =standard ; 25C =medium and 35C =high 2 Standard Deviation of the coefficients is shown in parenthesis. 3 where a, (a b)= regression coefficients and k = release rate. 4 NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.001.

PAGE 63

63 Table 2 9. Effect of temper a ture on total N released from five CRFs with time Total N Released (% applied) T ime (days) Cum. Incubation Temperature 1 7 14 28 42 56 81 112 140 180 (%) ------------Sulfur Coated Urea -------------1) 21C 1 33.9 8.0 8.3 5.3 4.9 5.7 6.4 2.2 1.9 76.5 2) 25C 30.0 10.9 10.2 3.2 4.2 3.7 3.4 1.4 3.4 70.4 3) 35C 43.6 4.6 5.5 2.9 1.4 4.2 6.0 2.1 3.8 74.0 1 vs. 2 NS ** ** NS 1 vs. 3 *** ** ** *** NS NS NS ** NS -----------Resin Coated NPK -----------1) 21C 26.2 9.8 13.7 15.7 7.3 12.8 5.2 2.4 1.6 94.7 2) 25C 34.8 19.2 13.7 9.8 6.3 6.9 2.8 1.2 0.0 94.6 3) 35 C 50.0 11.4 22.6 4. 8 3.9 6.6 2.2 0. 0 0.0 101.5 1 vs. 2 ** ** NS ** NS *** ** ** NS 1 vs. 3 *** NS *** ** *** ** *** *** *** NS -----------Polymer Sulfur Coated Urea -----------1) 21C 12.7 6.8 9.2 11.2 3.0 7.2 9.7 4.0 11.3 75.1 2) 25C 16. 8 13.6 12.1 10.8 6.3 7.5 2.4 3.5 1.2 74.1 3) 35C 25.0 11.3 19.5 4.5 4.7 7.9 3.3 2.1 3.2 81.5 1 vs. 2 NS ** NS NS *** NS *** NS ** NS 1 vs. 3 ** ** ** ** ** NS *** ** *** NS -----------Reactive Layer Coated Urea -----------1) 21C 14.6 14.4 19. 5 16.9 7.3 5.1 7.9 1.7 1.9 89.1 2) 25C 28.8 20.6 17. 1 3.9 2.7 5.8 2.2 1.1 1.1 83.2 3) 35C 39.1 15.3 22.5 3.7 3.0 4.0 1.5 0.7 0.6 90.3 1 vs. 2 ** *** *** NS *** ** NS 1 vs. 3 ** NS ** ** *** *** ** *** NS -----------Polyolefi n Coated NPK -----------1) 21C 2.8 1. 7 7.8 10.1 9.8 12.2 10.6 6.2 9.8 71.1 2) 25C 2.1 2.5 16.7 12.3 10.7 13.3 9.9 5.3 5.9 79.2 3) 35C 14.3 9.6 20.7 6.0 4.6 14.1 7.2 5.4 5.7 87.5 1 vs. 2 NS ** NS NS NS ** 1 vs. 3 *** *** ** ** ** *** ** *** ** 1 21C=standard; 25C=medium and 35C=high. 2 Single degree of freedom contrasts were generated using SAS GLM Proc. NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significan t P<0.0001.

PAGE 64

64 Table 2 10. Effect of temperature on total N released from three SRFs with time Total N Released (% applied) Time (days) Cum. Incubation Temperature 1 7 14 28 42 56 81 112 140 180 (%) ------------IBDU -------------1) 21C 13.9 18.7 20.5 8.7 6.1 3.5 5.0 1.6 1.2 79.2 2) 25C 14.2 22.7 28.0 2.7 6.1 3.6 1.3 1.4 0.4 80.0 3) 35C 32.5 16.1 18.3 3.5 3.1 3.0 1.4 0.5 0.0 78.8 1 vs. 2 NS NS *** *** NS NS *** NS ** NS 1 vs. 3 ** NS NS *** ** NS *** ** ** NS -----------Ureaform -----------1) 21C 14.6 1.1 3.9 1.5 1.4 1.4 2.7 1.6 3.7 31.9 2) 25C 16.4 2.9 1.3 1.3 1.3 2.3 2.5 2.4 3.4 33.8 3) 35C 20.3 2.7 5.5 1.7 1.3 5.9 5.6 2.7 4.8 50.4 1 vs. 2 NS ** ** NS ** NS *** ** NS 1 vs. 3 *** ** NS *** *** NS ** *** *** -----------Biosol id s -----------1) 21C 39.7 7.4 2.2 1.4 1.3 1.4 2.8 1.3 2.1 59.6 2) 25C 38.8 6.3 2.3 1.4 1.2 2.8 1.7 1.0 2.0 57.5 3) 35C 50.8 7.9 2.0 0.8 0.6 2.5 1.4 1.2 0.0 67.1 1 vs. 2 NS NS NS NS NS ** *** NS NS 1 vs. 3 NS NS ** *** ** *** NS *** 1 21C =standard; 25C=medium and 35C=high. 2 Single degree of freedom contrasts were generated using SAS GLM Proc. NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.0001.

PAGE 65

65 Table 2 11 Regression analysis of esti mated N release rate from five CRFs using the soil incubation method with four different soils. Soil origin 1 a (a b) k R 2 P value ------------Sulfur Coated Urea -------------1) Florida 78.191 48.597 0.0184 0.99 <0.0001 (1.184) 2 (1.240) (0.0017) 2) California 73.295 49.011 0.0305 0.96 <0.0001 (2.454) (4.541) (0.0023) 3)Pennsylvania 77.243 53.896 0.0281 0.98 <0.0001 (3.420) (1.231) (0.0033) 4) Iowa 78.008 68.458 0.0382 0.99 <0.0001 (3.451) (5.592) (0.0009) 1 vs 2 NS NS *** 1 vs 3 NS NS ** 1 vs 4 NS *** *** -----------Resin Coated NPK -----------1) Florida 97.041 83.775 0.0223 0.99 <0.0001 (7.918) (7.528) (0.0014) 2) California 95.441 77.633 0.0315 0.99 <0.0001 (5.127) (2.038) (0.0029) 3) Pennsylvania 92.155 80.457 0.0290 0.99 <0.0001 (0.484) (4.519) (0.0015) 4) Iowa 102.405 83.001 0.0337 0.99 <0.0001 (3.656) (4.326) (0.0028) 1 vs 2 NS NS ** 1 vs 3 NS NS ** 1 vs 4 NS NS *** -----------Polymer Sulfur Coated Urea -----------1) Florida 83.435 74.051 0.0105 0.99 <0.0001 (0.528) (1.112) (0.0005) 2) California 83.375 79.087 0.0201 0.99 <0.0001 (1.871) (1.645) (0.0012) 3) Pennsylvania 78.861 75.090 0.0282 0.98 <0.0001 (1.298) (1.168) (0.0020) 4) Iowa 77.327 75.102 0.0248 0 .99 <0.0001 (1.397) (2.577) (0.0036) 1 vs 2 NS ** ** 1 vs 3 ** NS *** 1 vs 4 ** NS ***

PAGE 66

66 Table 2 11. Continued. Soil origin 1 a (a b) k R 2 P value -----------Reactive Layer Coated Urea -----------1) Florida 88.248 91.134 0.0306 0.99 <0 .0001 (4.167) (2.690) (0.0006) 2) California 86.049 85.842 0.0361 0.99 <0.0001 (2.218) (1.358) (0.0020) 3) Pennsylvania 87.726 93.126 0.0385 0.99 <0.0001 (2.465) (4.665) (0.0005) 4) Iowa 97.276 102.697 0.0376 0.99 <0.0 001 (4.602) (5.327) (0.0003) 1 vs 2 NS NS *** 1 vs 3 NS NS *** 1 vs 4 ** ** *** -----------Polyolefin Coated NPK -----------1) Florida 96.136 101.153 0.0079 0.99 <0.0001 (1.932) (2.311) (0.0009) 2) California 90.023 97. 923 0.0107 0.99 <0.0001 (0.785) (1.162) (0.0007) 3) Pennsylvania 97.417 103.692 0.0123 0.99 <0.0001 (2.933) (2.805) (0.0011) 4) Iowa 91.707 101.510 0.0113 0.99 <0.0001 (1.385) (1.625) (0.0014) 1 vs 2 ** NS ** 1 vs 3 NS NS ** 1 vs 4 NS ** 1 1= Florida soil; 2= California Soil; 3= Pennsylvania soil and 4= Iowa soil. 2 Standard Deviation of the coefficients is shown in parenthesis. 3 where a, (a b)= regression coefficients and k = release rate. 4 N S = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.001.

PAGE 67

67 Table 2 12. Effect of soil type on total N released from five CRFs with time Total N Released (% applied) Time (days) Cum. Soil Origin 1 7 14 28 42 56 81 112 140 180 (%) ------------Sulfur Coated Urea -------------1) FL 33.9 8.0 8.3 5.3 4.9 5.7 6.4 2.2 1.9 76.5 2) CA 31.1 12.6 11.0 4.7 4.8 1.7 3.1 3.3 5.6 77.9 3) PA 30.6 12.1 12.8 4.8 4.7 3.8 4.9 3.1 2.5 79.3 4) IA 24.2 15.4 16.0 8.5 5.2 3.1 3.6 2.1 2.9 81.0 1 vs. 2 NS ** ** NS NS *** *** ** *** NS 1 vs. 3 NS ** *** NS NS ** ** NS NS 1 vs. 4 ** *** *** ** NS ** *** NS ** NS -----------Resin Coated NPK -----------1) FL 26.2 9.8 13.7 15.7 7.3 12.8 5.2 2.4 1.6 94.7 2) CA 32.4 14.0 17.4 9.7 10.3 3.9 5.3 1.9 1.4 96.1 3) PA 25.9 13.2 17.9 9.2 12.1 5.4 4.8 2.3 1.4 92.3 4) IA 34.2 20.1 16.5 9.1 10.3 5.8 4.1 2.3 1.3 103.6 1 vs. 2 ** ** ** *** NS NS NS NS 1 vs. 3 NS ** *** *** NS NS NS NS 1 vs. 4 ** *** ** ** *** NS NS NS -----------Polymer Sulfur Coated Urea -----------1) FL 12.7 6.8 9.2 11.2 3.0 7.2 9.7 4.0 11.3 75.1 2) CA 12.3 12.6 15.7 9.8 7.5 7.2 8.2 6.0 3.9 83.2 3) PA 13.5 13.7 18.4 7.5 8.5 5.6 3.9 4.6 3.9 79.6 4) IA 12.4 12.7 17.0 9.0 8.0 5.3 6.0 4.3 5.1 79 .7 1 vs. 2 NS ** *** *** NS NS *** *** *** 1 vs. 3 NS ** *** ** *** ** *** *** ** 1 vs. 4 NS ** *** ** *** *** ** NS *** ** -----------Reactive Layer Coated Urea -----------1) FL 14.6 14.4 19.5 16.9 7.3 5.1 7.9 1.7 1.9 89.1 2) CA 18.2 17.3 20 .7 10.0 8.5 3.7 5.1 2.4 1.9 88.1 3) PA 15.1 21.0 19.1 12.6 11.2 3.5 2.5 2.5 2.2 89.7 4) IA 17.7 19.9 24.3 12.5 12.3 4.9 3.0 2.4 1.8 98.7 1 vs. 2 NS NS ** ** *** ** NS NS 1 vs. 3 NS NS ** *** ** *** ** NS NS 1 vs. 4 ** ** *** NS *** ** NS **

PAGE 68

68 Table 2 12. Continued Soil Origin 1 7 14 28 42 56 81 112 140 180 Cum. (%) -----------Polyolefin Coated NPK -----------1) FL 2.8 1.7 7.8 10.1 9.8 12.2 10.6 6.2 9.8 71.1 2) CA 1.2 2.9 12.1 9.1 14.6 9.8 10.1 8.8 6.8 75.3 3) PA 5.0 2.5 14.4 11.3 17.2 10.0 9.3 8.4 8.4 86.5 4) IA 1.2 2.3 11.0 10.4 17.8 10.6 9.6 7.2 7.8 77.8 1 vs. 2 ** ** ** NS ** ** NS *** ** NS 1 vs. 3 *** *** NS ** NS ** NS *** 1 vs. 4 ** NS ** NS ** NS 1 FL, CA, PA, IA represents Florida, California, Pennsyl vania and Iowa respectively. 2 Single degree of freedom contrasts were generated using SAS GLM Proc. NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.0001.

PAGE 69

69 CHAPTER 3 OPTIMIZATION AND VALIDATION OF AN ACCELERATE D LAB EXTRACTION METHOD TO ESTIMATE NITROGEN RELEASE PATTERNS OF ENHANCED EFFICIENCY FERTILIZERS S everal methods to characterize the nutrient release properties of enhanced efficiency fertilizers (EEFs) have been developed during the last few decades The se procedures usually quantify nutrient release rates in either water or soil media depending on incubation time. Short term static or dynamic water incubation studies are commonly used for rapid estimation of nutrient release patterns of EEFs. The Katz me thod (AOAC 970.04) and the 7 d dissolution rate test developed by the Tennessee Valley Authority (TVA) were the first methods used to characterize the rate of N released in water. Several modifications of these procedures relative to incubation temperature and time have been studied since EEFs emerged commercially (Dai et al., 2008; Du et al., 2006; Novillo et al., 2001; Gambash et al., 1990; Blouin et al., 1971; Oertli and Lunt 1962). Despite the development of many procedures to characteriz e EEF materials no official standard method of analysis has been accepted for use by AOAC International S everal technologies have been tried to evaluate EEF nutrient release rates, but none have been shown to be accurate enough to verify label claims and material perfo rmance In the consumer perspective, it introduces confusion regarding choices when purchasing EEF and a lack of protection against ineffective products In 1994, a Controlled Release Fertilizer Task Force was established by the American Association of Pla nt Food Control Officials (AAPFCO) to address issues concerning the effective regulation and analysis of EEF materials (AAPFCO, 1995). As a result of task force

PAGE 70

70 efforts, a short term laboratory method based on the accelerated extraction of nutrients with t ime was developed (Sartain et al., 2004). Solid method validation is needed in order for this technique to be approved as an official laboratory procedure. Prior to any validation work, sufficient method development and optimiz ation should be performed to be reasonably confident that the Method o ptimization is the process of adjusting the experimental parameters of a method to find the levels that achieve the best possible response Usually, the classical, one variable at a time (OVAT) procedure is implemented to examine the major factors contribu ting to variability of a method during its development and optimization. The OVAT technique consists of var ying the levels of a given factor while keeping the othe r factors at standard levels in order to evaluate the effect of one specific factor on the method ( Dejaegher and Vander Heyden 2007) If a factor is found to have a major effect on method response, further method development is necessary. Ruggedness testi ng is also an important part in the development of a method. It is usually performed only after the method has been optimized because any successive changes will require re testing. Ruggedness testing evaluates the ability to reproduce a method in differen t laboratories or under different circumstances without the occurrence of unexpected differences in the obtained results ( Vander Heyden et al., 2001). Using the OVAT procedure to assess the effect of small changes in the experimental conditions on the vari ability of the method demands considerable experimental work and thus, the evaluation of only a small number of parameters is possible. Consequently, ruggedness testing is commonly based on partial factorial designs in

PAGE 71

71 order to balance the amount of inform ation obtained and the experimental work required. An economical approach based on fractional factorial designs has been described by Youden and Steiner (1975). A robust analytical method capable of producing repeatable nu trient release patterns from a bro ad range of EEF materials is required to gain acceptance as an official method by AOAC. Using this rapid technique, regulators and producers can monitor EEF efficiently in a laboratory setting. The objective s of this study were : 1) to i nvestigate the effec t of extraction temperature, fertilizer sample size and extraction time on the ability of an accelerated laboratory extraction method to estimate N P and K release rates of enhanced efficiency fertilizers ; and 2) to p erform ruggedness testing on the perf ormance of the accelerated lab extraction procedure to estimate N release patterns of EEF s using a fractional multifactorial design. The following hypotheses were tested in this study: 1) N P and K release rates increase with increasing extraction tempera ture and vice versa ; 2) s ample size affect s N, P and K release rates, increasing and decreasing with the high and low fertilizer sample size respectively; 3) r educing the extraction time reduce s the total amount of N, P and K extracted from the EEFs; and 4 ) s mall variations in several performance parameters of the method do not influence N, P and K release rates of the EEFs. Materials and Methods Standard Accelerated Lab Extraction Method The nutrient release profile of EEF materials was generated by accel erating their natural release mechanism in a laboratory setting. The standard method used to characterize the nutrient release properties of EEFs consist s of using a 0.2% citric acid solution to perform various solvent extraction procedures at increasing t emperatures on

PAGE 72

72 a non ground fertilizer sample E ach extraction is designed to remove and isolate nutrients that release or become available with time. A riffle is used to reduce the non ground fertilizer to a 30.0 g 1.0 g sample. The extraction procedure consist s of exposing the EEF to an accumulation and combination of extracts following the time and temperature sequence detailed in Table 3 1. An extra extraction (#5) consisting of 94 hr is only performed on an EEF if its longevity claimed by the manufac turer is more than 6 months. The main components of the extraction apparatus w ere the extraction columns and a water manifold. Eight vertical jacketed chromatography columns, enclosing inner columns of 2.5 cm x 30 cm where the EEF was placed, were used as the extraction columns (Figure 3 1). Polyester fiber (3 g 0.2 g) was placed 2 to 3 cm above the bottom of the column and near the top (~1cm) of the colu mn below the o ring of the cap to hold the fertilizer in place. A polytetrafluoroethylene (PTFE) rod ( 6 mm x 15 cm) was inserted inside the columns to avoid fertilizer caking. A 16 channel, reversible peristaltic pump was used to pass the extraction solution through the columns continuously for a maximum of 168 hr at a flow rate of 4.0 0.1 mL min 1 A w ater circulation manifold supplied by a centrifugal pump connected to a covered water bath capable of maintaining a maximum constant temperature of 60.0C 1.0 C for extended periods of time w as used to achieve the different temperatures in the system (Fig ure 3 2). Water circulated through the manifold and the chromatography columns at the same time The desired mean temperatures were obtained by preheating the water bath to several degrees (5 to 10C) above the desired temperature to account for initial he at exchange and temperature equilibration with the manifold and columns.

PAGE 73

73 Wrapping the water manifold and the chromatography columns with rubber pipe insulation was also required to maintain a stable temperature. Two in line, symmetrically placed thermomete rs were used to monitor incoming and outgoing temperatures to the manifold to assure stabilization of the system at the desired temperature. Roll clamps and flow monitors were also attached to the tubing connecting the water manifold to the columns to achi eve balanced flow and uniform temperature For each extraction, the water bath was adjusted to maintain the desired temperature in the columns. Erlenmeyer flasks (500 mL) were filled with 475 mL of 0.2% citric acid and this extraction solution and air wer e pumped from the flasks to the bottom of the columns through two different tubes (Figure 3 3). The solution circulated throughout the columns at a rate of 4.0 0.1 mL min 1 for a specific time at a specific temperature according to the extraction sequenc e. The flasks were swirled occasionally to mix the solution during the extraction. At the end of each extraction period, the flow was reversed and air was pumped into the columns from the top to collect the sample at the bottom in to the flasks (Figure 3 4) After the columns were emptied of liquid a ir was pumped for 1 min to assure complete transfer of solution. The solution was then cool ed to room temperature and diluted to volume (500 mL) with extraction solution. A 250 mL volume of extract was placed in a graduated cylinder and 5.0 m L of c upric s ulfate stabilizing solution (20 g Cu S O 4 5H 2 O per L of 1 : 1 HCl acid) were added for sample preservation. An aliquot was taken from this extract and frozen for later nutrient analysis. For the first extraction, the remain ing 250 mL of extract was discarded and the second extraction was started with fresh extraction solution. However, for the consecutive extractions, the remain ing 250 mL of extract was saved to be used in the subsequent extraction, i.e. for

PAGE 74

74 extra ctions 3 through 5, only 225 mL of fresh extraction solution was added to the flask to bring the total starting volume to 475 mL. The liquid extracts from each extraction sequence were then analyzed for total N using an Antek 9000 N analyzer (PAC Company, Houston, TX). Phosphorus concentration in the leachates was determined with a Bran Leubbe Technicon Autoanalizer I I (Seal Analytical, Mequon, WI ) as automated The potassium (K) concentration was analyzed at the University of Florida Analytical Research Laboratory following USEPA method 200.7 (USEPA, 199 4) using an Inductively Coupled Plasma Spectrophotometer (ICP). Results for each extraction were presented as cumulative percent age of total nutrient released. Extraction #1 was considered to represent the readily available water soluble fraction of the EE F. Optimization of the Accelerated Lab Extraction Method A fully developed and optimized method is necessary to obtain successful validation and method utilization. Various experiments were performed to evaluate the effect of temperature, fertilizer sample size and extraction time on the performance of the accelerated lab extraction method to measure the nutrient release profile of EEFs. Preliminary t emperature s tudy Temperature is the most important factor that can potentially affect nutrient release patt erns from coated EEFs. T hus, additional experiments were conducted to evaluate this factor. Prior to the series of experiment s performed to optimize this method a preliminary temperature study was done to identify the temperature range to be used in the o ptimization studies. T wo polyolefin coated (POC) fertilizers were used to evaluate the effect of temperature on N release rate: POC N PK Type A 18 6 18 with a

PAGE 75

75 manufacturer defined release period of 100 d, and POC NPK Type B 18 6 18, with a 270 d release pe riod. These two fertilizers had the same nutrient core but different coating thickness allowing evaluation of the membrane thickness effect on N release characteristics. Four experiments were conducted following the standard accelerated lab extraction met hod described above at the four extraction temperatures shown in Table 3 2. Method o ptimization s tudies The OVAT procedure was used to evaluate the effect of temperature, fertilizer sample size and extraction time on N, P and K release patterns from fou r EEFs. In each experiment, only the factor under evaluation was varied at predetermined interval ranges while keeping all other experimental parameters at the levels described in the operating procedures of the standard accelerated lab extraction method ( Table 3 1) The effect of each factor was individually evaluated by comparing the results obtained after varying one factor with that of the experiment with all factors at standard levels. Four fertilizer materials including different EEF technologies and defined release periods were used in five experiments (Table 3 3). The first experiment characteriz ed N, P and K release patterns of the selected EEFs for 74 hr under the conditions of the standard accelerated lab extraction method. Based on the results f rom the previous temperature study, only the medium extraction temperature (maximum 55C) described in Table 3 2 was further evaluated using the broader range of EEF materials. Two more experiments were also performed to evaluate the influence of fertilize r sample size on nutrient release rates. A 10% variation in sample size was followed resulting in one low sample size (27 g) and one high sample size (33 g) experiment. Since one of the main purposes of this method is to

PAGE 76

76 develop the nutrient release prof ile of EEFs in a short period of time, only a reduced extraction time was considered in the optimization study. Roughly a 10% decrease in extraction time was applied only to the 2 nd and 3 rd extractions in order to investigate the effect of reduced extracti on time in the short term (2 nd extraction) and long term (3 rd extraction). The reduced time extraction sequence used in the experiment was: Extraction #1 2 :00 hr.; Extraction #2 1:45 hr.; Extraction #3 18:00 hr.; and Extraction #4 50 :00 hr. Statistica l a nalysis F erti lizer material treatments were arranged in a randomized complete block design with three replications Fine particle size IBDU (0.5 to 1.0 mm) was included in each run as a standard reference material due to its high dependence on water to release N and its low water solubility. For the statistical analysis, the first extraction was omitted in all experiments since it wa s considered to be the water soluble fraction of the EEF T hus only the slow release fraction of the release curve was use d for mean comparison. Each experiment conducted to evaluate any factor (temperature, sample size, or extraction time) w as compared with the standard method at normal levels Data were statistically analyzed using the Statistical Analysis System (SAS versi on 9.0; SAS Institute, Cary, NC). Means were s eparat ed using the general linear model procedure Ruggedness Testing of the Accelerated Lab Extraction Method The purpose of ruggedness testing is to evaluate the effect of small variations in several parame ters on the results of the method using a limited number of experiments. The ruggedness of the accelerated lab extraction method to measure N, P, and K release rates from EEF was evaluated using the Youden & Steiner ruggedness test

PAGE 77

77 (Youden and Steiner 197 5) This design provides a set of multifactorial experiments that permit varying up to seven factors simultaneously and require only a total of eight experiments Factors representing quantitative and qualitative variables were selected from the operationa l conditions of the standard method. The seven factors that were evaluated in this ruggedness trial with the variations for each factor are presented in Table 3 4. All other parameters were identical to those in the standard accelerated lab extraction proc edure previously described. Five EEF materials replicated three times each were used to assess the ruggedness of this method (Table 3 5). These EEFs were selected with the intent of incorporating a wide range of EEF technologies. The responses determined i n this robustness test were cumulative percentage N, P, and K released at each extraction time (2; 4; 24; 74 hr) for each individual EEF. The fractional factorial design used to evaluate the selected factors is shown in Table 3 6. In this table, the origin al and modified method parameter values were expressed by upper case (+) and lower case letters ( ) respectively, i.e., changing a factor from value A to value a. This design is considered well balanced because each factor is used the same number of times as (+) and ( ) levels. In each subset of experiments, all other factors were also used the same number of times as (+) and ( ) levels (Alvares Ribeiro and Machado, 1997). For example, in a subset of experiments that had all (+) levels such as 1 to 4 for fa ctor A, all other parameters also occurred twice as the (+) level and twice as the ( ) level. The effect of each factor (D x ) was calculated from the eight experimental results following the Youden & Steiner test that uses the equations below. By using the se equations, the main effects of all other parameters were canceled out when the average

PAGE 78

78 difference between the two levels of a factor was calculated. Thus, each equation (D x ) measured individually the effect of changing factor X from one level to the oth er on the N, P and K release rates of the selected EEFs. D a = (R 1 + R 2 + R 3 + R 4 R 5 R 6 R 7 R 8 ) D b = (R 1 + R 2 R 3 R 4 + R 5 + R 6 R 7 R 8 ) D c = (R 1 R 2 + R 3 R 4 + R 5 R 6 + R 7 R 8 ) D d = (R 1 + R 2 R 3 R 4 R 5 R 6 + R 7 + R 8 ) D e = (R 1 R 2 + R 3 R 4 R 5 + R 6 R 7 + R 8 ) D f = (R 1 R 2 R 3 + R 4 + R 5 R 6 R 7 + R 8 ) D g = (R 1 R 2 R 3 + R 4 R 5 + R 6 + R 7 R 8 ) After calculating the main effect of each factor, the significance of the D x values was evaluated statistically and graphic normal probability plots, respectively. A full explanation of the e stimation of the experimental error from the distribution of the effects using the algorithm of Dong can be found in Vander Heyden et al. ( 2001 ) Table 3 7 shows the parameters that were estimated and used to statistical ly interpret the significan ce of the effects. All parameters were calculated independently for each EEF at each extraction time. An effect ( D x ) that is between the margin of error ( ME ) and the simultaneous margin of error ( SME ) values is considered to be possibly signifi cant and an effect that is above the SME is considered to be signifi cant However, the ME limit was used in this study as the decision criterion due to its recommended use for all effects calculated from a ruggedn ess test ( Vander Heyden et al., 2001). Half normal probability plots were used as a graphical tool to identify significant effects of the sele cted factors on the method A complete description of the construction of half normal probability plots can be found in Nijhuis et al. ( 1999 ) Half normal plots were created for each individual EEF by ranking all the effects derived from its

PAGE 79

79 corresponding ruggedness test (D x ) according to an increasing absolute effect size and then plotting the value s of t hat sequence against the corresponding rankits The rankit values for the fractional factorial design with eight experiments used in this study were extracted from Vander Heyden et al. ( 2001 ) (Table 3 8). In a half normal proba bility plot, non significant effects tend to fall on a straight line through zero, while a deviation from it indicates that the effect involved is probably significant. The detection of significant effects in these plots is done visually and thus it can be very subjective. Therefore, the decision limits to statistically interpret these plots. The significance of an effect was determined following the same criterion described above in the statistical interpretation of the effects. R esults and Discussion Optimization of the Accelerated Lab Extraction Method In the preliminary temperature study, the rate of N release was greatly influenced by temperature (Table 3 9 ). C umulative percent age N released from both EEFs increased with temperature. Increasing the maximum temperature from 50 to 55C and 60C almost tripled the total cumulative percent age N released from both fertilizer materials. Regardless of temperature, N release rate s from POC NPK Type B were slower than those from POC NPK Type A. Although both are polyolefin coated materials a slower N release rate was expected from POC NPK Type B since it has a thicker coating to provide a longer nutrient release period. At the low temperature sequence (max imum 50C), N released from either EEF was less than 35% of the total indicating that the temperature was too low to accelerate the N release rate in a short period of time. When comparing standard with medium

PAGE 80

80 temperature, no sig nificant difference in N release rates was generally found for POC NPK Type A, while the opposite trend was observed for POC NPK Type B. In contrast, when comparing standard with high temperature, there was a significant difference in N release rates for b oth EEFs. However, these temperature effects were more marked on POC NPK Type B, resulting in almost 30 to 40% more total N released at the medium and high temperatures, respectively, compared with the standard temperature. These findings demonstrated that N release rates from thickly coated EEFs are more influenced by temperature changes than N release from thinly coated EEFs. Despite the difference s in N release rates with increasing temperature Figure 3 5 suggest s that N release patterns from both EEFs at the medium and high temperatures were comparable with those obtained at the standard temperature. In addition, at these temperatures, it was possible to differentiate N release patterns of fertilizer products that had the same coating technology but dif ferent thicknesses Therefore, it was recommended to further evaluate the maximum extraction temperature in the 55 to 60C range using a broader selection of EEF materials in order to determine a temperature sequence that provides reproducible, well differ entiated N release curves while maintaining an easy to operate system. Increased wear and tear of the tubing and consequently greater chances for leaks in the system were experienced at 60C. The effect of temperature, fertilizer sample size, and extracti on time on N release rate are shown in Table 3 10. Overall, none of the factors had an effect on total cumulative percentage N released from any EEF in 74 hr of extraction. However, a slight effect of temperature and extraction time on N release rates fro m polymer coated NPK Type B was observed throughout the extraction sequence. Total cumulative N

PAGE 81

81 released from this fertilizer was 23% less at the medium temperature compared with the standard. Likewise, 42 and 27% less N was recovered from the second and t hird extractions, respectively, when extraction time was reduced compared with the standard method. These results were expected since polymer coated NPK Type B has a thick coating ( greater longevity) therefore a higher temperature and longer extraction ti me were needed to generate its N release profile using the rapid extraction method. A graphic representation of the effects of temperature, fertilizer sample size, and extraction time on N release patterns from the four EEFs is show n in Figure 3 6. Despite the slight differences in N release rates at some extraction times, each EEF followed similar N release patterns across experiments. Only temperature highly influenced the cumulative percentage P released from both EEFs (Table 3 11). Reducing the extracti on temperature to the medium sequence (max imum 55 C ) generally decreased P release rate at all extraction times for both EEFs. At the medium temperature (max imum 55 C ), the total cumulative percentage P released was reduced almost by half for both fertiliz ers compared with the standard temperature (max imum 60 C ). This result indicated that increasing the temperature to 60 C during the last 50 hr of extraction greatly increased the amount of P released. Similarly, only temperature had an effect on K release rate during 74 hr of extraction for both EEFs. As shown in Table 3 12, small differences were observed relative to the total quantity of K recovered at each extraction time with respect to temperature change. However, the cumulative percentage K released f rom polymer coated NPK Type B was 32% lower at the medium temperature compared with the standard method. The release rate of nutrients from NPK coated fertilizers is expected to be a more complex process compared with release

PAGE 82

82 from a single nutrient coated fertilizer. Numerous studies have reported different trends in the differential release of N, P, and K from polymer coated fertilizers ( Oertli and Lunt 1962 ; Huett and Gogel 2000 ; Du et al. (2006) In general, across the different polymer coated fertili zers studied, the nutrient release rate ranking was: N>>P >K Across all temperatures, the data depicted in Figure 3 7 illustrate that generally similar total quantities of P and K were released from both EEFs, with polymer coated NPK Type A releasing 35 to 50%, 30 to 36% and 40 to 47% more N, P, and K respectively, compared with Type B due to its thinner coating. For both EEFs, N release rate was greatest, followed by P and K. The influence of the core fertilizer solubility on the release rate of N, P, and K through the polymer coating was obvious. T hus the ranking of nutrient release, N>P>K, confirmed reasonably well with the solubility of their corresponding fertilizer components (water solubility at 20 C is 192, 37 and 32 g/100mL for ammonium nitrate (AN ), monoammonium phosphate (MAP) and potassium nitrate (KN) respectively). At the standard temperature, total cumulative N released was 25 to 27% and 30 to 36% higher than the total cumulative P and K releas ed, respectively (Figure 3 7a and c). In contrast at the medium temperature (max imum 55 C), the difference in total cumulative P and K released compared with N was 44 to 64% and 41 to 48%, respectively (Figure 3 7b and d). It was noteworthy that these differences increased with decreasing temperature; h owever, this effect was greater for P. The results indicated t h at temperature exerted substantial influence on the release of P and K ions which was likely due to the increased solubility of these compounds (MAP and KN) and higher diffusion rate s as the t emperature increased Similar findings were presented by

PAGE 83

83 Tomaszewska and Jarosiewicz (2002) and Zhang et al. (1999) on the effect of temperature on nutrient solubility and diffusion rates of coated NPK and MAP fertilizer f ormulations Ruggedness Testing of the Accelerated Lab Extraction Method A combination of graphical and statistical approaches was used to identify significant effects of the selected factors on the performance of the accelerated lab extraction method. The results obtained from the eight e xperiments evaluating N, P, and K release rates are presented in Tables 3 13, 3 14 and 3 15 respectively. The values were expressed in cumulative percentage nutrient released at each extraction time during a 74 hr time period. The factor effects ( D x ), stan dard deviations ( S 0, S 1 ), and ME and SME; ) were calculated independently for each EEF material at each extraction time The algorithm of Dong was used to statistically interpret the effect of the factors on N, P, and K release rate s. Tables 3 16, 3 17 and 3 18 show the e ffect of the seven factors ( D a through D g ) on the cumulative percentage s of N, P, and K released respectively, and their corresponding critical effect values (ME) T he effect values ( D x ) for most EEFs me t the accepta nce cri terion ( D x < ME ), indicat ing that the cumulative percentage s of N, P, and K released were generally not significantly influenced by the variation of the factors from the standard condition s For N, a bsolute D x values higher than criterion val ues were observed only with RLCU and IBDU for the A and B factors (Table 3 16). For RLCU, temperature influenced the cumulative percentage N released after 4 hr of extraction. The positive D b value indicated that decreasing the temperature in the second extraction from 50 to 40 C decreased the cumulative percentage N released from RLCU from 7.6 to 5.9. A

PAGE 84

84 decrease in N released with decreasing temperature was expected because the diffusion coefficient of RLCU and the solubility of urea depend on temperature. Despite the fact that temperature change caused a 22% decrease in N released compared with the standard conditions in the second extraction, this effect was isolated and not o bserved with any of the other coated fertilizers studied. Therefore, varying the temperature parameter in the predetermined range (Table 3 4) was not considered to significantly affect the performance of the method to measure N release rates from coated EE Fs. For IBDU, the flow rate of the extract ion solution and extraction temperature had an effect on the cumulative percentage N released at the third extraction, resulting in positive D a and D b values of 5.5 and 5.8, respectively. Thus, compared with the st andard conditions, reducing the flow rate of the extract solution to 3.5 mL/min and the temperature of the second extraction to 40 C resulted in a decrease of cumulative percentage N released from 104.5 to 99.0 and 104.7 to 98.9 respectively. Since the rat e of IBDU dissolution is affected by temperature and moisture, the effects of flow rate and temperature on N released from IBDU were expected. However, these factors only caused a minor reduction in N released (~5%) compared with the standard conditions. I n addition, since IBDU is used as a standard reference material, its purpose is to assess the capability of the method to measure N release rates from EEFs. Thus, a complete recovery of the applied N as IBDU is required after 74 hr of extraction. On averag e, a recovery of about 100% of the applied N was already obtained by the third extraction regardless the experiment (Table 3 13). Therefore, the effect of factors A and B on IBDU were considered to have little influence on method results.

PAGE 85

85 Only two EEFs (re sin coated NPK and polyolefin coated NPK) were used to test P and K ruggedness because the other three fertilizer materials contained N only. None of the factors had an effect on the cumulative percentage P released at any extraction time, while factor A ( flow rate of the extract solution) produced a large effect on the total cumulative percentage K released from resin coated NPK (Tables 3 18 and 3 19). This effect resulted in a positive D a value of 28.6, suggesting that a decrease in the flow rate of the e xtract solution reduced the total cumulative percentage K released after 74 hr of extraction. Although a significant reduction in percentage K released was only obtained in the last extraction, the same trend was generally seen at all extraction times (Tab le 3 15, experiments 6, 7, and 8). Overall, compared with the standard condition (4 mL/min), 35 to 50% less K was released at any extraction time when a flow rate of 3.5 mL/min was used. Lower K release rates at a flow rate of 3.5 mL/min probably resulted from a lower concentration gradient between the fertilizer granules and the surrounding solution, since less solution circulated through the columns and thus a higher concentration of K was maintained around the granules compared with the standard conditio n. Half normal probability plots were algo rithm to graphically interp ret the results and identify significant effects The rankit values in Table 3 8 were used to build half normal plots for all EEF s at each extraction t ime However, only the plots that showed significant effects are presented. Figures 3 8 and 3 9 show the effects of the selected factors on the cumulative percentage N released from RLCU and IBDU, respectively, while Figure 3 10 shows the effects on cumula tive percentage K released from resin coated NPK None of the factors had a significant effect on P

PAGE 86

86 released from any EEF. The results described by the half normal probability plots agreed with the results obtained from the statistical interpretation using the algorithm of Dong. These half normal probability plots clearly indica ted that most effects were m ore or less equal to random erro r since they mostly fell on a straight line through the origin. Cumulative percentage N released from IBDU and RLCU was si gnificantly influenced by factors A & B and B only, respectively (Figures 3 8 and 3 9). Likewise, factor B affected cumulative percentage K released from resin coated NPK (Figure 3 10). These factors evidently deviated from the straight line, and their eff ect values (D x ) were greater than their corresponding ME values. Thus, they were determined to have a significant effect in the interval specified by the r ugged ness test.

PAGE 87

87 Figure 3 1. Extraction apparatus with eight jacketed chromatography colum ns. Figure 3 2.Schematic diagram of water manifold used in the extraction apparatus.

PAGE 88

88 Figure 3 3.Schematic diagram of the extraction phase. Figure 3 4.Schematic diagram of the collection phase.

PAGE 89

89 Figure 3 5.Nitrogen released from POC NPK Typ e A and Type B in 168 hr as affected by four extraction temperature sequences

PAGE 90

90 Figure 3 6.Effect of temperature, fertilizer sample size, and extraction time on N released from ureaform, PCU, PC NPK Type A, and PC NPK Type B in 74 hr compared with the standard accelerated lab extraction method.

PAGE 91

91 Figure 3 7. Release of N, P, and K from two EEFs in 74 hr using the accelerated lab extraction method (a: PC NPK Type A at st andard tempera ture; b: PC NPK Type A at medium temperature; c: PC NPK Type B at standard temperature; d: PC NPK Type B at medium temperature )

PAGE 92

92 Figure 3 8. Half normal probability plot for seven effects on percentage N released from RLCU with identification of the crit ical effects ME and SME. Figure 3 9 Half normal probability plot for seven effects on percentage N released from IBDU with identification of the critical effects ME and SME.

PAGE 93

93 Figure 3 10 Half normal probability plot for seven effects on percentage K released from resin coated NPK with identification of the critical effects ME and SME.

PAGE 94

94 Table 3 1. Sequence of extraction times and temperatures for the standard accelerated lab extraction method. Extraction Number Extraction Time (hr) Temperat ure (C) 1 2 25 2 2 50 3 20 55 4 50 60 Table 3 2. Sequence of extraction times and temperatures used in the preliminary temperature study. Extraction Number Extraction Time (hr) Standard Temperature (C) Low Temperature (C) Medium Temperature (C) H igh Temperature (C) 1 2 25 25 25 25 2 2 50 50 55 60 3 20 55 50 55 60 4 50 60 50 55 60 5 94 60 50 55 60 Table 3 3. Description of the enhanced efficiency fertilize r used in the optimization studies Source Formulation (N P 2 O 5 K 2 O) Release duration ( months) 1 Principle source 2 N P 2 O 5 K 2 O Polymer Coated Urea 44 0 0 1 2 WSON ------Ureaform 38 0 0 6 9 70% WIN 18%SAWSN 12% WSON ------Polymer Coated NPK Type A 15 7 15 4 AN,MAP KN M AP KN Polymer Coated NPK Type B 15 7 15 8 AN,MAP KN MAP KN 1 Approximate at 21C soil temperature. 2 WSON= water soluble organic N (urea); AN= ammonium nitrate; MAP=monoammonium phosphate; KN= potassium nitrate; WIN=water insoluble N ; SAW SN=slowly available water soluble N

PAGE 95

95 Table 3 4 Factors and their levels for ruggedness testing of the accelerated lab extraction method. Factor Nominal Conditions Modified Conditions Flow rate of extract solution A= 4.0 mL/min a= 3.5 mL/min Extraction temperature B= 25C; 50C; 55C; 60C; b= 25C; 40C; 55C; 60C; Concentration of extract solution C= 0.20% c= 0.25% Addition of stabilizer solution D= no d= yes Amount of extraction solution E= 475 mL e= 450 mL N form in standards F= ammonium nitrat e f= urea Flow meters of water manifold G= all open g= 4 fast & 4 slow Table 3 5 Enhanced effi ciency fertilizers used in the ruggedness testing of the accelerated lab method N Source Formulation (N P 2 O 5 K 2 O) Release duration (months) 1 Principle source 2 N P 2 O 5 K 2 O Resin Coated NPK 19 6 12 3 4 AN, AP AP,CP KS Polymer Sulfur Coated Urea 37 0 0 6 PSCU ------Polyolefin Coated NPK 18 6 18 6 AN, AP, KN AP,CP KN Reactive Layer Coated Urea 43 0 0 2 3 PCU -----IBDU 31 0 0 2 3 WSON ------1 Approximate at 21C soil temperature, as specified by the manufacturer. 2 WSON= water soluble organic N (urea); AN= ammonium nitrate;AP=ammonium phosphate; CP=calcium phosphate; KS= potassium sulfate; PSCU=polymer s ulfur coated urea; PCU=polymer coated urea and KN=potassium nitrate.

PAGE 96

96 Table 3 6 Experiment Number Factor Combination Experimental Response A/a B/b C/c D/c E/e F/f G/g 1 A B C D E F G R 1 2 A B c D e f g R 2 3 A b C d E f g R 3 4 A b c d e F G R 4 5 a B C d e F g R 5 6 a B c d E f G R 6 7 a b C D e f G R 7 8 a b c D E F g R 8 Table 3 7 logarithm Parameter Formula Condition Initial estimate of the standard error Where is the valu e of effect Final estimate of the standard error For all Where number of Margin of error Where Simultaneous margin of er ror Where Table 3 8 Rankits used to draw half normal probability plots using eight experiments Effect Rankit Value 1 0.09 2 0.27 3 0.46 4 0.66 5 0.90 6 1.21 7 1.71

PAGE 97

97 Table 3 9 Effect of four temperature sequences on total cumulative %N released from two EEFs using the standard accelerated lab extraction method. Time (hr) Treatments (Cumulative %N released) 1 Contrasts 2 (1) Std. (2) Low TP (3) Medium TP (4) High TP 1 vs 2 1 vs 3 1 vs 4 ------------POC NPK Type A -------------2 0.0 0.0 0.0 0.0 4 0.0 0.0 0.4 0.4 NS 24 18.6 7.9 23.7 30.8 *** *** 74 60.7 21.8 62.1 70.0 *** NS ** 168 86.1 33.8 85.9 95.9 *** NS ** -----------POC NPK Type B ----------2 0.0 0.0 0.0 0.0 4 0.0 0.0 0.0 0.2 NS NS *** 24 3.3 2.6 6.5 10.1 NS *** *** 74 28.2 12.4 35.1 40.1 *** ** *** 168 49.8 23.0 63.1 68.1 *** *** *** 1 Treatments: 1= standard temperature (max. 60 C ) ; 2=low temperature (max. 5 0 C); 3= mediu m temperature (max. 55C); 4= high temperature (max. 60 C) 2 NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.001.

PAGE 98

98 Table 3 10 Effect of temperature, fertilizer sample size, and extraction time on tota l N released from four EEFs using the accelerated lab extraction method. Time (hr) Treatments (Cumulative %N released) 1 Contrasts 2 (1) Std. (2) Medium TP (3) Low SS (4) High SS (5) Red. Time 1 vs 2 1 vs 3 1 vs 4 1 vs 5 ------------Polymer Coated Urea -------------2 0.7 0.8 1.0 0.8 1.0 4 6.2 2.2 3.5 4.2 4.9 *** ** ** NS 24 62.8 55.7 74.0 62.7 74.3 NS NS 74 84.3 75.3 90.9 81.9 93.7 NS NS NS NS ----------------Ureaform ---------------2 15.6 14.2 16.3 15.0 15.8 4 24.0 22.0 24. 4 21.2 23.9 NS NS 24 28.3 26.2 28.4 27.6 28.7 NS NS NS NS 74 31.5 28.5 31.1 29.5 33.0 NS NS NS NS -------------Polymer Coated NPK Type A -----------2 0.0 0.0 0.0 0.0 0.0 4 2.1 1.5 1.7 2.2 2.3 NS NS NS 24 28.3 27.2 37.7 31.3 29.6 NS NS NS 74 55.6 57.4 57.7 57.8 61.5 NS NS NS NS -------------Polymer Coated NPK Type B -----------2 0.3 0.3 0.0 0.0 0.1 4 1.9 1.4 1.2 1.6 1.1 ** ** *** 24 9.2 8.3 9.6 10.7 6.7 NS NS NS 74 36.2 28.0 31.2 37.7 32.0 NS NS NS 1 Treatmen ts: 1= standard method; 2= medium temperature (max. 55C); 3= low sample size (27 g); 4=high sample size (33 g); 5= reduced time (2:00; 1:45; 18:00; 50:00 hr). 2 NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.00 1.

PAGE 99

99 Table 3 11 Effect of temperature, fertilizer sample size, and extraction time on total P released from two EEFs using the accelerated lab extraction method. Time (hr) Treatments (Cumulative %P released) 1 Contrasts 2 (1) Std. (2) Medium TP (3) Low S S (4) High SS (5) Red. Time 1 vs 2 1 vs 3 1 vs 4 1 vs 5 -------------Polymer Coated NPK Type A -----------2 0.0 0.0 0.0 0.0 0.0 4 1.0 0.3 0.7 0.7 0.8 ** NS NS 24 14.0 11.4 17.6 15.5 14.6 NS NS NS NS 74 41.7 20.8 37.6 37.8 39.6 *** NS NS NS -------------Polymer Coated NPK Type B -----------2 0.1 0.1 0.0 0.1 0.1 4 1.0 0.6 0.5 0.7 0.7 ** ** 24 6.4 4.9 5.7 5.9 5.4 NS NS NS 74 26.5 14.6 27.2 25.9 27.1 *** NS NS NS 1 Treatments: 1= standard method; 2= medium temperature (max. 55C) ; 3= low sample size (27 g); 4=high sample size (33 g); 5= reduced time (2:00; 1:45; 18:00; 50:00 hr). 2 NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.001.

PAGE 100

100 Table 3 12 Effect of temperature, f ertilizer sample size, and extraction time on total K released from two EEFs using the accelerated lab extraction method. Time (hr) Treatments (Cumulative %K released) 1 Contrasts 2 (1) Std. (2) Medium TP (3) Low SS (4) High SS (5) Red. Time 1 vs 2 1 vs 3 1 vs 4 1 vs 5 -------------Polymer Coated NPK Type A -----------2 0.1 0.1 0.1 0.2 0.1 4 0.8 1.4 1.7 1.6 1.4 ** ** ** ** 24 14.8 8.9 18.4 18.6 15.1 NS NS NS 74 38.9 33.7 45.6 41.8 31.4 NS NS NS -------------Polymer Coated NPK Type B ----------2 0.2 0.1 0.2 0.2 0.2 4 0.9 1.1 1.1 1.0 1.2 NS NS NS 24 5.4 4.4 6.7 6.9 6.7 NS NS NS 74 23.3 15.8 28.7 25.8 19.7 NS NS NS 1 Treatments: 1= standard method; 2= medium temperature (max. 55C); 3= low sample size (27 g); 4=high sample si ze (33 g); 5= reduced time (2:00; 1:45; 18:00; 50:00 hr). 2 NS = not significant, *= significant P<0.05, **= significant P<0.01 and ***= significant P<0.001.

PAGE 101

101 Table 3 13 Results of eight ruggedness experiments on the cumulative %N release d for five EEFs using the accelerated lab extraction method. Cumulative %N Released Extraction Time (hr) Exp 1 R 1 Exp 2 R 2 Exp 3 R 3 Exp 4 R 4 Exp 5 R 5 Exp 6 R 6 Exp 7 R 7 Exp 8 R 8 ------------Resin coated NPK -----------2 11.4 11.4 11.1 10.9 9.1 9.5 15.8 9.5 4 22.5 23.4 22.0 21.6 27.8 21.4 32.7 19.4 24 47.1 49.8 45.3 49.8 50.5 42.8 57.3 46.7 74 96.1 94.5 97.1 97.4 88.3 100.6 93.4 93.2 -----------Polymer sulfur coated urea -------2 4.7 5.2 5.7 4.3 3.0 4.8 3.6 4.7 4 16.0 16.5 14.2 24 .2 14.3 15.5 11.2 12.1 24 40.6 35.8 35.4 34.5 33.3 36.2 37.3 33.7 74 51.0 45.7 49.6 48.2 61.3 50.4 51.6 45.8 --------Polyolefin coated NPK ------2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 4 0.5 0.4 0.2 0.2 0.3 0.3 0.3 0.4 24 13.0 11.0 6.7 12.1 26 .7 13.6 12.5 13.0 74 53.6 50.7 57.7 52.5 65.9 53.0 62.2 54.0 --------Reactive layer coated urea ------2 2.5 2.0 2.3 2.3 2.8 2.5 2.3 2.1 4 8.1 6.9 6.2 6.3 8.0 7.4 6.0 5.2 24 47.7 42.1 33.7 39.3 42.6 41.9 41.1 36.6 74 85.5 77.6 88.4 75.1 8 1.7 81.0 77.4 75.3 ---------IBDU ---------2 9.8 10.2 10.1 9.2 8.1 8.9 8.1 9.2 4 58.6 55.2 35.9 48.0 54.5 45.7 39.0 33.2 24 108.4 105.1 103.0 101.4 100.9 104.2 94.4 96.6 74 110.6 107.5 109.5 106.1 104.7 109.2 100.1 102.7

PAGE 102

102 Table 3 14 Res ults of eight ruggedness experiments on the cumulative %P released for two EEFs using the accelerated lab extraction method. Cumulative %P Released Extraction Time (hr) Exp 1 R 1 Exp 2 R 2 Exp 3 R 3 Exp 4 R 4 Exp 5 R 5 Exp 6 R 6 Exp 7 R 7 Exp 8 R 8 ------------Resin coated NPK -----------2 4.7 11.6 5.7 6.8 13.6 3.1 3.9 4.3 4 10.2 22.1 11.0 11.0 23.7 6.0 9.3 7.5 24 21.5 42.4 28.8 44.8 38.2 26.8 27.2 33.5 74 58.6 62.4 67.6 67.3 59.5 56.8 58.2 46.0 --------Polyol efin coated NPK ------2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 24 2.6 2.6 1.2 2.1 4.1 1.9 1.6 2.3 74 11.6 12.5 9.2 8.3 16.9 8.8 10.5 6.0 Table 3 15 Results of eight ruggedness experiments on the cumulative %K released for two EEFs using the accelerated lab extraction method. Cumulative %K Released Extraction Time (hr) Exp 1 R 1 Exp 2 R 2 Exp 3 R 3 Exp 4 R 4 Exp 5 R 5 Exp 6 R 6 Exp 7 R 7 Exp 8 R 8 ------------Resin coated NPK -----------2 10 .0 11.4 11.0 10.2 17.1 5.3 5.0 5.6 4 19.7 22.7 20.3 18.2 29.1 10.9 9.9 11.0 24 40.7 50.7 50.7 48.0 47.3 24.4 29.5 26.5 74 95.2 98.3 88.9 90.9 77.1 58.8 65.3 57.9 --------Polyolefin coated NPK ------2 0.0 0.0 0.0 0.1 0.1 0.0 0.1 0.0 4 0.2 0.1 0.1 0.1 0.2 0.1 0.2 0.1 24 3.3 3.7 2.0 2.5 6.1 2.2 7.8 2.6 74 21.5 19.2 11.5 9.5 13.7 8.8 26.8 7.4

PAGE 103

103 Table 3 16 Effects of seven factors on the cumulative %N released for five EEFs using Extra ction Time (hr) D a D b D c D d D e D f D g Critical Effect (ME) 1 ------------Resin coated NPK -----------2 0.2 1.5 1.5 1.9 1.5 1.7 1.6 3.6 4 2.9 0.2 4.8 1.3 5.1 2.0 1.4 7.2 24 1.3 2.2 2.8 3.1 6.4 0.3 1.2 7.3 74 2.4 0.4 2.7 1.6 3.4 2.7 3.6 6.1 -----------Polymer sulfur coated urea -------2 1.0 0.2 0.5 0.1 1.0 0.7 0.3 1.5 4 4.4 0.2 3.1 3.1 2.1 2.3 2.5 6.6 24 1.4 1.3 1.6 2.0 1.2 0.6 2.6 3.9 74 3.7 3.3 5.9 3.8 2.5 2.2 0.3 8.2 --------Polyolefin coated NPK ------2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4 0.0 0.1 0.0 0.1 0.1 0.1 0.0 0.2 24 5.7 5.0 2.3 2.4 4.0 5.3 1.5 9.6 74 5.2 0.8 7.3 2.1 3.2 0.6 1.7 8.9 --------Reactive layer coated urea ------2 0.1 0.2 0.2 0.3 0.0 0.1 0.1 0.4 4 0.2 1.7* 0.6 0.4 0.0 0.3 0.4 0.9 24 0.1 5.9 1.3 2.5 1.3 1.9 3.7 7.1 74 2.8 2.4 6.0 2.6 4.6 1.7 1.0 8.1 ---------IBDU ---------2 1.2 0.1 0.4 0.3 0.6 0.3 0.4 1.4 4 6.3 14.5 1.4 0.5 5.8 4.6 3.1 15.9 24 5.5* 5.8* 0.2 1.2 2.6 0.2 0.7 3.4 74 4.3 3.4 0.2 2.2 3.4 0 .6 0.4 6.1 1 Margin of error, calculated using ME= t (1 S 1 ; ( =0.05). *Significant to ME: x ME, ( =0.05)

PAGE 104

104 Table 3 17 Effects of seven factors on the cumulative %P released for two EEFs using terion. Extraction Time (hr) D a D b D c D d D e D f D g Critical Effect (ME) 1 ------------Resin coated NPK -----------2 1.0 3.1 0.5 1.2 4.5 1.3 4.1 6.4 4 2.0 5.8 1.9 0.6 7.8 0.9 6.9 9.5 24 3.0 1.3 8.0 3.5 10.5 3.2 5.6 13.8 74 8.9 0.5 2.8 6. 5 4.6 3.4 1.4 11.4 --------Polyolefin coated NPK ------2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 24 0.4 1.0 0.2 0.1 0.6 0.9 0.5 1.5 74 0.2 4.0 3.2 0.6 3.2 0.5 1.3 5.5 1 Margin of error, calculated using ME= t (1 s 1 ; ( =0.05). Table 3 18 Effects of seven factors on the cumulative %K released for two EEFs using Extraction Time (hr) D a D b D c D d D e D f D g Critical Effect (ME) 1 ------------Resin coated NPK -----------2 2.4 3.0 2.6 2.9 3.0 2.6 3.7 6.9 4 5.0 5.7 4.1 3.8 4.5 3.6 6.1 11.3 24 15.6 2.1 4.6 5.7 8.3 1.8 8.2 18.8 74 28.6* 6.6 5.2 0.3 7.7 2.4 3.0 23.9 --------Polyolefin coated NPK ------2 0.0 0.0 0.0 0. 0 0.0 0.0 0.0 0.0 4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 24 1.8 0.1 2.0 1.1 2.5 0.3 0.3 3.5 74 1.2 2.0 7.1 7.8 5.0 3.5 3.7 11.6 1 Margin of error, calculated using ME= t (1 s 1 ; ( =0.05).

PAGE 105

105 CHAPTER 4 STATISTICAL CORRELATION OF THE SOIL INCUBATION AND THE ACCELERATED LAB EXTRACTION METHODS TO ESTIMATE NITROGEN RELEASE RATES OF ENHANCED EFFICIENCY FERTILIZERS Currently, a wide range of enhanced efficiency fertilizers (EEF) is produced and distributed in the United States, Canada, China, Japan, Europe and Israel for different purposes in the agriculture and horticulture sectors (Trenkel 2010). However, there is no universally accepted o fficial method to verify the nutrient release patterns claimed by the manufacturers of these EEFs. A Controlled Release Task Force established in 1994 by AAPFCO was formed to meet regulatory needs and address effective analysis of EEFs As a result, a new method involving an accelerated extraction of nutrients in a laboratory setting was developed to generate data that could be correlated with the real time nutrient release profile of EEFs The accelerated lab extraction method exposes the EEF to a sequence of increasingly aggressive temperatures in order to accelerate its natural release mechanism and develop its N release profile in a short period of time (74 hr) However, the N release mechanism of some EEF products depend s at least partially on biologica l activity. Consequently some concern was expressed by the task force regarding the applicability of an accelerated extraction method to characterize these types of EEF products As a result, the soil incubation method wa s developed as a sister method th at although not accelerated, is biologically active and can be conducted under controlle d conditions. Therefore, there was a need to assess the ability of the accelerated lab extraction method to produce data that could be correlate d with the release of N under the controlled, biologically active soil incubation method.

PAGE 106

106 The objective of this study was to develop a prediction model that can use data from the accelerated lab extraction method to predict the 180 d N release curve for a specific EEF product. I t was hypothe sized that the accelerated lab extraction method generate s data that can be correlated with the N release data from the soil incubation method in order to predict the N release curve for a specific EEF. Materials and Methods Nonlinear Regress ion No n linear regression techniques we re used to establish the correlation between the data generated from the accelerated lab extraction method and the soil incubation method and to develop the prediction model All analyses were performed using the Sta tistical Analysis System (SAS version 9.0; SAS Institute, Cary, NC) Performing a nonlinear least squares regression is a slightly different process compared with a linear least squares regression. The derivatives for a linear model do not depend on any pa rameters, but in a nonlinear model the derivative matrix is a function of at least one element of the estimation matrix. For this reason, SAS PROC NLIN fits a nonlinear regression model using a least squares iterative process while the nonlinear model is approximated by a series of least squares linear models with user supplied initial values (SAS/STAT 9.2 User's Guide). Each iteration is an attempt to further improve the resulting parameter estimates. SAS PROC NLIN includes four different methods to perf orm nonlinear regression: Gaussian Newton (default), Newton, Marquardt, and Steepest Descent/Gradient. T o determine the most appropriate method for this study, several nonlinear regression models were fitted using all four iterative methods with identical user supplied initial values. Results indicated that the method used to estimat e the update vector had no

PAGE 107

107 effect on the resulting estimation matrix. Thus the robust G aussian Newton method was used for all nonlinear estimation s. Prediction Model s A two st ep analysis process was used to correlate soil incubation data with results from the accelerated lab extraction. The first step fit ted nonlinear regression curves to the N release data from long term soil incubation These data had an exponential like stru cture that is typical of N release curves. The nonlinear equation [ was assumed to be the functional form of the soil incubation N release curve (Sartain et al., 2004). The estimates of the release constants a, b and k were saved to serve as dependent variable values in the second step. Step 1: (1) Where, N(t) = cumulative percentage N released with time t = time, in d a = maximum amount of N released (asymptote of the curve) b = intercept k = rate, with units of percentage N released/d In the second step, the valu es of the release constants ( a, b and k ) and the four extraction values ( E XT 1, E XT 2, E XT 3 and E XT 4 ) generated by the accelerated lab extraction procedure were used to fit equations of a multiple regression model in which E1, E2, E3 results were estimates of the coefficients ( 0 4 0 4 0 4 ) in step 2 Step 2:

PAGE 108

108 (2) (3) (4) Where, = first, second, third and fourth accelerated lab extraction values respectively. Then for each fertilizer material, the mean extraction v alues we re obtained and inserted into the fitted equations 5, 6 and 7 to pre dict the values of the release constants (a, b and k) Finally, the predicted values of the release constants we re used to create the esti mated N release curve for a 180 d period (Equation 8) Prediction: (5) (6) (7) (8) Where, = estimates of a, b, and k = coefficient estimates of 0 4 0 4 and 0 4 = mean extraction v alues from E XT 1, E XT 2, E XT 3 E XT 4 Data issues Clearly, the ideal situation was to fit a model for the N release curve (equation 1) separately for each EEF then use those constants ( a, b and k ) as the dependent variables to fit a non linear regressi on curve using the accelerated lab extraction method dat a as independent variables However, with at most four replications of the accelerated lab extraction method for each EEF it was not possible to estimate each

PAGE 109

109 coefficient of the nonlin ear regression curve by individual EEF (equation s 2 3, and 4 ) because the number of observations must be more than th e number of parameters ( 0 4 0 4 and 0 4 ) that are being estimat ed in order to calculate a unique parameter estimate. Analysis of variance (ANOVA) was performed on the model parameter estimates (a, b and k) for the soil incubation method (Equation 1) to determine if they differe d by fertilizer material. If the parameter estimates were not different, stratification of the model by fertilizer material would not be necessary, and as a result the sample size would increase However, with a P value < 0.0001, the parameter estim ates were different and thus, the model needed to be stratified by fertilizer material. Two groupings were proposed: Group A was defined based on the N release mechanism of the EEFs, while Group B was defined based on statistics only. Group A divided the E EFs into three subgroups that included: 1) RLC u rea SCU, resin coated NPK, polyolefin coated NPK and Poly S ; 2) IBDU, urea form, m ethylene u reas A, B and C; and 3) b lend A and B. Group B had four subgroups: 1) RLC u rea IBDU, SCU, resin coated NPK; 2) pol yolefin coated NPK Poly S and methylene urea B; 3) urea form, m ethylene u rea A, m ethylene u rea C; and 4) b lend A and B. The optimal grouping was identified using ANOVA for the stratified estimates and was chosen according to the group ing that produced the largest statistically significant difference between subgroups and had the higher R 2 for the estimates. Table 4 1 displays the statistics that resulted from the ANOVA on the corresponding N release model coefficients (Equation 1) to determine differences between grouping definitions. For both groups A and B, statistically significant differences between subgroups were

PAGE 110

110 consistently observed for a, (a b), and k with a P value < 0.0001. Furthermore, the R 2 values were generally lower for the group A coeffici ents compared with the statistics of group B. Although R 2 for group A was lower for the a and k coefficients, it was slightly larger for the (a b) coefficient s However, th e increase d R 2 for the (a b) coefficients using group A was not worth the lower R 2 v alues for the a and k coefficients compared with group B. Therefore, group B definition appeared more able to detect differences between subgroups, indicating that estimates of the N release curve parameters using the accelerated extraction data would impr ove A lthough this group was defined based on s tatistics, the subgroups did coincide with some N release mechanisms of the EEFs Choosing group B also implied that errors associated with these estimates would be smaller compared with group A and associate d prediction intervals would be tighter. Once the two step process need ed to create unique e stimates was determined, the linear regression model assumptions of norma lity, homogenous variances, and independence were evaluated N ormality o f the data was not checked, as it was not necessary to estimat e regression parameters or to partition the total variation of th e data (Rawlings et al ., 1998 ). I nd ependence of the data was assumed based on how the data were processed and collected. Assessing whether the data ha d hom ogenous variances was necessary, as estimation can be affected if this assumption was not met. After graphing the res iduals and the predicted values for each variable, there was no evidence of heterogeneity within the grouped data. Th erefore, the li near regression model assumptions were met. Zero extraction values Polyolefin coated NPK was the only fertilizer that had no measurable N released by the first extraction. It was questioned whether these zero values actually help ed to

PAGE 111

111 demonstrate, or provi de d any information about the nutrient release profile of this fertilizer material. Basically, it was essential to determine if using these zero values account for this var iation all the analyses were performed using the original accelerated lab extraction data (including the zero extraction values) and then the same analyses were conducted using only the non zero accelerated lab extraction values. The results for each anal ysis were compared and the data set that gave a better prediction for N release was the final data set used for each analysis method. Twelve EEFs with at most four replications each, were subjected to both the soil incubation method and the accelerated la b extraction method. A description of the EEF p roducts is presented in Table 4 2 The resulting cumulative percent age N release d data from both methods were s tatistically analyzed and used to generate the three prediction models detailed below. One replica tion of each EEF which was selected randomly, was not used in the model fitting process but was saved for model validation procedures. Two step nonlinear regression method with grouping The two step process shown in equations 1 4 was applied to the group ed data to obtain coefficient estimates for 0 4 0 4 0 4 The predicted values for a b and k ( were then used to calculate the predicted N release data through 180 d [ This method was performed using the original accelerated lab extraction values and also using only non zero extraction values. For each of these data sets, the two step regression analysis us ed the first two accelerated lab extraction values and the first three accelerat ed lab extraction values for each replication. The

PAGE 112

112 data were also analyzed using all four accelerated lab extraction values but this analysis was possible only with the original data due to over p arameterization issues. P rincipal component analysis with g rouping Principal component analysis (PCA) was another proposed analysis method because it was reasonable to speculate that the accelerated lab extraction valu es could be considered multiple i ndepende n t variables that essentially measure d nearly the same t hing. Highly correlated variables can affect estimation of the coefficients and can produce very large standard errors for these estimates. PCA shrinks the number of independent variables so that the estimation matrix is not nearly singular, resulting in m ore reliable estimates and smaller corresponding standard errors (Joliffe 2002 ). This process uses the available accelerated lab extraction data to create composite variables called principal components T hese re defined variables then take the place of t he extraction values and are used as the explanatory variables in any analysis. The number of principal components created from the acc elerated lab extraction values wa s determined by studying the scree plot s for the corresponding stratified data (Figure 4 1). The s cree plots indicate d that at least the first two principal components should be included in the model since the line bega a fter the second eigenvalue Associated, non stratified output indicated that 95% of the variability was explained by the first two principal components. Once defined, the first two principal components were used in the multiple linear regressions to est imate the coefficients ( 0 4 0 4 0 4 ) for the N release curve in step 2 ( equations 2 4 ) shown above. Then, the se results were used to obtain the predicted N release curve [ using equations 5 6, and 7 for each fertilizer material. This method was

PAGE 113

113 performed using the original four accelerated lab extraction values and also only the non zero accelerated lab extraction values. Likewise, this process was performed using two and three principal components f or each accelerated lab extraction data set Two s tep m ethod b ased on PCA without g rouping As previousl y mentioned, any model used to predict the N release curve should be fitted by fertilizer material. However, the number of replicates for the accelerated lab extraction data for each EEF w as insufficient to uniquely estimate the fi ve coefficients found in step 2 A third prediction method that can be described as a hybrid of the two previously proposed analysis methods but did not use any grouping of the fertilizer materials was also evalua ted. This method followed the same process as the PCA method that was based on the principle of correlated explanatory variables However, instead of redefining the explanatory variables into principal components as was done in the second method a subset of the accelerated lab extraction values was used as the explanatory variables in the multiple linear regressions (equations 2, 3, and 4) and thus, material specific estimates for the coefficients were calculated Since non stratified data were used in thi s prediction method, it was necessary to optimize the number of explanatory variables that could be used for each fertilizer material but would still result in unique parameter estimates. Based on the scree plot analysis described in the PCA method, the fi rst two accelerated lab extraction values explained 95% of the variation in the un grouped data, indicating that using the first two extraction values should be sufficient for a model. Unlike the other fertilizer materials, which had three replications of t he accelerated lab extraction data available, ureaform and Poly S had four replications available. For these two EEFs only the prediction model used the first three extraction values. For all

PAGE 114

114 other EEF materials, using the first three extraction values w a s impossible as only three replications were available, so only the first two extraction values were used. The end results were material specific estimates of the coefficients shown in equations 2 3, and 4 (step 2). After the coefficients were estimated, the next step was to compute the predicted values for a b and k based on the fitted models for each material T hen those predicted values were used to calculate the predicted N release data separately for each EEF material, as outlined in step 1. Results and Discussion The nonlinear regression [ was fitted to the mean N release data obtained from the soil incubation method separately for each EEF (step 1). An R 2 statistic, equal to the sum of squared deviations about the fitted curve divided by the sum of squared deviations for the mean, was calculated for each EEF to determine the goodness of fit of the nonlinear regression to the N release data. The coefficients for the fitted N release curve and the corresponding R 2 are prese nted in Table 4 3 for each EEF. The R 2 values were greater than 0.99 for most EEFs except for ureaform and methylene urea A that had an R 2 value of 0.98. Therefore, the nonlinear equation fitted the observed N release data very well. For all prediction met hods, results were similar between the original accelerated lab extraction values and the data that used only the non zero accelerated lab e xtraction values. Since the predictions were very close, us e of the original accelerated lab extraction values for p rediction was preferred because the prediction model could be uniformly created regardless of the EEF material s included in each grouping. Figures 4 2 through 4 7 display the predicted N release curves for ea ch permutation of data sets and number of explan atory variables for the two step nonlinear regression method and

PAGE 115

115 the PCA with grouping method. The mean soil incubation N release data for each EEF were graphed along with each prediction model to compare the predicted N release curves with the N release d ata obtained from the soil incubation. For each prediction method, the permut ation of data set and number of explanatory variables that produced N release curves closer to the soil incubation N release data were chosen. Then, a final prediction model was s elected from the three prediction methods F igures 4 2 through 4 4 show the predicted N release curves using the two step nonlinear regression method with two, three, and four extraction values. Predictions for most EEF s were similar regardless of whether two, three, or all four accelerated lab extraction values were used H owever the prediction model using four accelerated lab extraction values w as closest to the mean soil incubation N release values for methylene urea A, resin coated NPK, and methylene u rea C Therefore, the prediction model using all four original accelerated lab extraction data values was selected as the final model for this prediction method. Figures 4 5 through 4 7 display the predicted N release curves using the PCA with grouping me thod with two and three principal components. For most EEFs the PCA method produced similar predictions using either two or three principal components, all of which were created from all four original accelerated lab extraction values. T he prediction mode l using two principal components w as closest to the mean soil incubation N release values only for RLC urea and poly S. H owever polyolefin coated NPK and blend B presented the same pattern for the prediction model using three principal components. Since t he scree plot s indicated that two principal components were appropriate and they explained at least 90 % of the variability of the data for all

PAGE 116

116 groupings (95% using non stratified data), the chosen model was the PCA with grouping prediction method using two principal components. Final Prediction Model A final prediction model was selected from the three different prediction methods evaluated throughout this study. Figures 4 8 though 4 10 show a graphic comparison of the predicted N release curves obtained fr om the three selected methods for each EEF. The third model t he PCA like non stratified prediction method produced results that were typically further from the mean soil incubation N release values compared with the other prediction methods. This trend w as clearly seen with methylene urea A, RLC urea, SCU, and blend B (Figures 4 8 and 4 10). For the rest of the EEFs, the predicted N release curves were very close to the other two prediction methods that are easier to use Therefore, the PCA like non strat ified prediction method was not further evaluated The l eave one out cross validation (LOOCV) method was u s ed to evaluate the remaining two prediction models and to determine the best method to predict the cumulative percent age N release rates of the twelv e EEFs The LOOCV values for each EEF ma terial and total were calculated with the two step nonlinear regression method and with the PCA with grouping method (Table 4 4 ). Overall, a smaller cross validation value indicates better model fit. In this case, si nce the prediction methods were not chosen for individual EEFs but for the best prediction method as a whole considering all EEF materials, the total LOOCV value from each prediction method was used as the LOOCV result to reference. As shown in T able 4 3 the total LOOCV value for the two step nonlinear regression prediction method was the smallest of the two methods and therefore this method was selected as the final prediction model.

PAGE 117

117 The predicted N release curves using the final method along with the fi tted N release curves using the soil incubation N release data are presented in Figures 4 11 through 4 13. The predictive ability of this final method was assessed by calculating a R 2 type statistic equal to (SS1 SS2)/SS1, where SS1 equals the sum of squar ed differences between points on the fitted N release curve and the overall mean, and SS2 equals the sum of squared differences between points on the fitted N release curve and points on the predicted N release curve. This pseudo R 2 measures how much bette r the predicted N release curve agrees with the fitted N release curve than a straight line through the mean agrees with the fitted N release curve. An R 2 value equal to 1 indicates that the predicted N release curve is exactly the same as the fitted curve and a value of 0 indicates that the predicted N release curve fits no better than a horizontal line. Tabl e 4 5 presents the estimated coefficients for the final prediction model with the corresponding R 2 values for each EEF. L arger differences between th e predicted and the fitted N release curves were found for ureaform, poly S, and polyolefin coated NPK (Figures 4 11 through 4 13). These results were further confirmed by the lower R 2 values of 0.76, 0.89, and 0.94 for these fertilizers, respectively, com pared with the other EEFs. The R 2 values for t he rest of the EEF materials were greater than 0.97, which suggested that the accelerated lab extraction method was capable of predicting N release rates with time from these EEFs with very high accuracy. Error Estimation Error terms for the two step nonlinear regression method were difficult to calculate b ecause th is prediction model is a two step process. Bootstrapping of the data was used as a method to estimate the error terms associated with the coefficient estimates.

PAGE 118

118 Bootstr apping is the practice of estimating properties of an estimator by measuring those properties when sampling from an approximating distribution. For this analysis, the empirical distribution of the observed data was used as the approximat ing distribution. By p erforming the bootstrapping in this manner, it was inherently assume d that the observations we re from an independent and identically distributed population (Efron and Tibshirani 1993 ). This assumption was possible because within each stratum, all the replications came from the same EEF The resampling was performed using sample sizes equal to the size of the observed data set, each of which was obtained by random sampling with replacement from the original data set. Smoothed bootstrap ping uses the same bootstrapping method with a small amount of random age N released. For this analysis N 2 ) was used as random noise, where and n is the associated sample size This random noise was added to each bootstrap sample to obtain a T his smoothed bootstrapping method, stratified by EEF mat erial was performed using 10,000 simulations The bootstrapping results were used to create 90% prediction intervals for each EEF. Figures 4 14 through 4 16 show the predicted N release curve with the associated prediction interval for each EEF. Predictio n intervals for ureaform, RLC urea, IBDU, resin coated NPK, Poly S and polyolefin coated NPK were very narrow. In general, the prediction intervals for each EEF could be misleading. Only three replications for each EEF were available t o perform the bootstr apping that used 10,000 simulations per fertilizer material. With the small sample size for each EEF it wa s possible that the variation captured in the data wa s insufficient to produce appropriate

PAGE 119

119 prediction intervals, indicati ng that the error estimates could be biased (Efron and Tibshirani 1993; Efron, 1982 ). However, it was even difficult to determine if this bias existed, and if so, by how much because of the small amount of data available

PAGE 120

120 Figure 4 1. Scree plots for four group s of EEFs (1: RLC u rea SCU IBDU and r esin c oated NPK ; 2: Poly S p olyolefin c oated NPK and m ethylene u rea B ; 3: u reaform m ethylene u rea A and m ethylene u rea C ; 4: b lend A and b lend B ).

PAGE 121

121 Figure 4 2. Predicted N release curves using the two step nonlinear regression method with 2, 3, and 4 extraction values for u reaform methylene urea A, RLC u rea and SCU

PAGE 122

122 Figure 4 3. Predicted N release curves using the two step nonlinear regression method with 2, 3, and 4 extraction values for IBDU, resin coated NPK, blend A, and Poly S.

PAGE 123

123 Figure 4 4. Predicted N release curves using the two step nonlinear regression method with two, three, and four extraction values for polyolefin coated NPK, methylene urea A, methylene urea C, and blend B.

PAGE 124

124 Figure 4 5. Predicted N release curves using the PCA with grouping method with two and three principle components for u reaform methylene urea A, RLC u rea and SCU

PAGE 125

125 Figure 4 6. Predicted N release curves using the PCA with grouping method with two and three principle components for IBDU, resin coated NPK, blend A, and Poly S.

PAGE 126

126 Figure 4 7. Predicted N release curves using the PCA with grouping method with two and three principle components for polyolefin coated NPK, methylene urea A, methylene urea C, and blend B.

PAGE 127

127 Figure 4 8. Predicted N release curves using the three selected prediction methods for u reaform methylene urea A, RLC u rea and SCU (Model 1: two step nonlinear regression method; Model 2: PCA with grouping method; and Model 3: PCA like non stratified method ).

PAGE 128

128 Figure 4 9. Predicted N release curves using the three selected prediction methods for IBDU, resin coated NPK, blend A, and Poly S (Model 1: two step nonlinear regression method; Model 2: PCA with grouping method; and Model 3: PCA l ike non stratified method ).

PAGE 129

129 Figure 4 10. Predicted N release curves using the three selected prediction methods for polyolefin coated NPK, methylene urea A, methylene urea C, and blend B (Model 1: two step nonlinear regression method; Model 2: PCA with grouping method; and Model 3: PCA like non stratified method ).

PAGE 130

130 Figure 4 11. Predicted N release curves using the final prediction method (two step nonlinear regression) for u reaform methylene urea A, RLC u rea and SCU

PAGE 131

131 Figure 4 12. Predicted N r elease curves using the final prediction method (two step nonlinear regression) for IBDU, resin coated NPK, blend A, and Poly S.

PAGE 132

132 Figure 4 13. Predicted N release curves using the final prediction method (two step nonlinear regression) for polyolefin co ated NPK, methylene urea A, methylene urea C, and blend B.

PAGE 133

133 Figure 4 14. Predicted N release curves using the final method with the corresponding 90% prediction intervals for u reaform methylene urea A, RLC u rea and SCU

PAGE 134

134 Figure 4 15. Predicted N re lease curves using the final method with the corresponding 90% prediction intervals for IBDU, resin coated NPK, blend A, and Poly S.

PAGE 135

135 Figure 4 16. Predicted N release curves using the final method with the corresponding 90% prediction intervals for poly olefin coated NPK, methylene urea A, methylene urea C, and blend B.

PAGE 136

136 Table 4 1. Statistics from ANOVA of fitted N release curve coefficients by group Group A 1 Group B 2 a (a b) k a (a b) k Sum of Squared Error 15889.05 8404.87 0.0073 11224.151 12303. 665 0.0030 R 2 0.623 0.893 0.624 0.7334 0.843 0.845 P value Grouping <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 1 Group A: 1) RLCU, SCU, resin coated NPK, polyolefin coated NPK, and Poly S; 2) IBDU, ureaform, methylene ureas A, B and C; and 3) blend A and B. 2 Group B: 1) RLCU, IBDU, SCU, resin coated NPK; 2) polyolefin coated NPK, Poly S, and methylene urea B; 3) ureaform, methylene ureas A and C; and 4) blend A and B.

PAGE 137

137 Table 4 2 E nhanced efficiency fertilizer s used for the statistical correlation of the methodologies N Source Formulation (N P 2 O 5 K 2 O) Release duration (months) 1 Principle source 2 N P 2 O 5 K 2 O Ureaform 38 0 0 6 9 70% WIN 18%SAWSN 12% WSON ------Methylene Urea A 40 0 0 4 35% WIN 50%SAWSN 15% WSON ------Re active Layer Coated Urea 43 0 0 2 3 PCU ------Sulfur Coated Urea 39 0 0 1 2 WSON ----IBDU 31 0 0 2 3 WSON ------Resin Coated NPK 19 6 12 3 4 AN, AP AP,CP KS Blend A 20 2 20 -60%WIN(Ureaform) KN CP KN Polymer Sulfur Coated Urea 37 0 0 6 PSCU ------Polyolefin Coated NPK 18 6 18 6 AN, AP, KN AP,CP KN Methylene Urea B 40 0 0 -40%WIN ------Methylene Urea C 30 0 0 -30%WIN ------Blend B 33 0 4 -54%WIN(Ureaform) 14%SAWSN 32% WSON ---KS 1 Approxim ate at 21C soil temperature, as specified by the manufacturer. 2 WSON= water soluble organic N (urea); AN= ammonium nitrate; AP=ammonium phosphate; CP=calcium phosphate; KS= potassium sulfate; PSCU=polymer sulfur coated urea; PCU=polymer coated urea; KN=p otassium nitrate WIN=water insoluble N ; SAWSN=slowly available water soluble N

PAGE 138

138 Table 4 3. Regression coefficients and R 2 values for the fitted soil incubation N release data for twelve EEFs. N Source a (a b) k R 2 Ureaform 47 4 6 0 2 9. 41 0 0.00 52 0.9 8 (16.036) 1 (14.898) (0.0033) Methylene Urea A 5 8 684 32. 402 0.016 9 0. 9 8 (5.237) (1.005) (0.0022) RLC Urea 89 749 88.925 0.033 9 0.99 (12.532) (12.271) (0.0072) SCU 96.93 5 84.5 3 8 0.038 7 0. 99 (6.245) (2.037) (0.0033) IBDU 8 9. 3 5 1 9 8. 55 9 0.0 395 0. 9 9 (13.220) (21.895) (0.012) Resin Coated NPK 103.3 6 2 86 5 71 0.02 42 0. 99 (6.982) (9.105) (0.0058) Blend A 60.1 17 23.4 80 0.0074 0. 99 (3.763) (3.690) (0.0007) Poly S 8 2 .6 52 74. 6 15 0.01 67 0.9 9 (7.533) (7.373) (0.0064 Polyol efin Coated NPK 8 1 .595 8 8 412 0.0132 0. 9 9 (12.966) (13.480) (0.0022) Methylene Urea B 94. 2 13 59.2 3 9 0.031 7 0. 99 (7.729) (5.362) (0.0041) Methylene Urea C 67. 424 2 9. 0 22 0.012 8 0. 9 9 (3.634) (0.641) (0.0022) Blend B 91.4 99 4 3 002 0.0142 0. 99 ( 4.098) (8.060) (0.0039) 1 Standard deviation of the fitted coefficients is shown in parenthesis.

PAGE 139

139 Table 4 4 L eave one out cross validation results for the two step nonlinear r egression method and the PCA with grouping method for twelve EEFs N Source Non Linear Regression PCA 1 with Grouping Ureaform 33.35 10.34 Methylene Urea A 1.15 1.68 RLC Urea 11.33 4.93 SCU 0.82 5.94 IBDU 29.5 0 38.45 Resin Coated NPK 3.49 6.97 Blend A 2.72 6.69 Poly S 83.06 187.16 Polyolefin Coated NPK 49.03 136.09 Methylene Urea B 8.06 5.32 Methylene Urea C 12.29 76.35 Blend B 33.93 55.28 Total 268.73 535.21 1 PCA= Principal component analysis

PAGE 140

140 Table 4 5 Regression coefficients and R 2 values for the predicted N release curves using the final selected m ethod ( two step nonlinear regression ) for twelve EEFs N Source a (a b) k R 2 Ureaform 37.126 22.741 0.0084 0. 76 (5.519) 1 (0.823) (0.0023) Methylene Urea A 59.030 32.457 0.0168 0. 99 (0.824) (0.358) (0.0007) RLC Urea 90.183 91.215 0.0 294 0. 98 (0.448) (0.385) (0.0014) SCU 96.955 84.586 0.0386 0. 99 (0.552) (0.929) (0.0006) IBDU 90.456 87.982 0.0270 0. 98 (2.285) (0.931) (0.0008) Resin Coated NPK 103.376 87.761 0.0231 0. 99 (0.825) (1.561) (0.0011) Blend A 6 0.132 23.479 0.0074 0. 99 (2.104) (3.877) (0.0006) Poly S 83.464 74.115 0.0105 0. 89 (2.167) (2.135) (0.0017) Polyolefin Coated NPK 76.161 81.64 5 0.0110 0. 94 (2.676) (1.070) (0.0014) Methylene Urea B 94.329 59.294 0.0316 0. 99 (3.459) (3.692) (0.0012) Methylene Urea C 67.703 28.605 0.0129 0. 97 (4.337) (1.659) (0.0007) Blend B 91.482 42.968 0.0142 0. 99 (4.617) (7.981) (0.0040) 1 Standard deviation of the predicted coefficients is shown in parenthesis.

PAGE 141

141 CHAPTER 5 CONCLUSIONS Long Term Soil Incubation Method The objective of this study was to e valuate the effect of changes in sand/soil ratio, incubation temperature and soil type on the N release rates of EEFs with time using a 180 d soil incubation method. The following hypotheses were tested in this study: 1) N release rates of the SRFs increase with increasing amount of soil present in the incubation mixture, and vice versa; 2) N release rates of the CRFs are not affected by the amount of soil present in the incubation mixture; 3) N re lease rates of the EEFs increase with increasing temperatures; and 4) soil texture does not influence N release rates of the EEFs. The gradual lowering of leachate pH for the fertilizer treatments during 180 d of incubation was in small part due to the co ntinued addition of 0.01% citric acid and in large part due to nitrification occurring in the incubation media. This result suggested that the system was functioning natural ly, with conditions favorable for microbial activity. The decline in leachate pH to about 3.5 at the end of the experiment did not appear to inhibit nitrification because N was still recovered during the last leaching events. Likewise, t h e mild effect of the citric acid on the system pH was further confirmed by the reduction of less than one unit in leachate pH of the non amended soil columns until the end of the incubation period In general, volatile ammonia N was detected only through the first 7 d of incubation. This observation was likely caused by the initial leachate pH of 8.0 from most EEFs due to the hydrolysis of the urea to ammonium. The decrease of pH in the subsequent leachates due to nitrification of the ammonium was most likely the reason no more volatile ammonia was detected through the termination of the

PAGE 142

142 experimen ts. About 96% of the N applied as water soluble AN was recovered after 14 d which was an indicator of the suitability of the system for its intended purpose The R 2 values for all the equations were close to unity, and all relationships except for the bio solid s material were statistically significant at P < 0.0001. This result indicated that the equations provided a good approximation of the N release rate (% of applied) with time. Regardless of the study, all EEFs followed similar N release patterns. In g eneral, sand/soil ratio had an effect on the rate of N release of the EEFs, which varied depending on the type of fertilizer. For the SRFs (IBDU, ureaform and the biosolids), N release rates and the total cumulative N released through 180 d decreased with the low soil sample size (45 g soil) and increased with the high soil sample size (180 g soil) compared with the standard soil sample size (90 g soil). This sand/soil ratio effect on the SRFs was more marked with the columns receiving the high soil sample size. These results suggested that the slightly higher moisture content and microbial activity of this incubation system influenced the rate of N release from the SRFs. For the CRFs, N release rates were only influenced by the high soil sample size, produc ing a higher initial N release during the first 14 d but similar total cumulative N released through 180 d of incubation, compared with the standard treatment. Temperature had the greatest influence on N release rates from EEFs. For C RFs the initial N rel ease rates and the percent age N release per day ( k ) increas ed as temperature increased. This temperature effect was more pronounced at 35C. However, the total cumulative N released was similar at any temperature. For SRFs a n increase in temperature from 25 to 35C increased the initial N release rates and the total cumulative N released, and almost doubled the percent age N release d per day

PAGE 143

143 from ureaform and biosolid s This temperature effect on the EEFs led to the conclu sion that increasing the temperatur e from 2 1 to 35C accelerated the diffusion rate of the CRFs, the dissolution and hydrolysis rate of IBDU, and the microbial degradation of ureaform and biosolid s. N release rates from all EEF s were influenced by soils that var ied in texture from sandy to loam y Soil texture had an influence on the percent age N release per day ( k ) which increased follow ing the order: Iowa >California>Pennsylvania>Florida. These resu lt s demonstrated that N was released faster from columns inoculated with soils high in clay an d/or organic C content and slower with sandy and silt loam soils. The soil incubation column leaching technique was demonstrated to be a robust and reliable method for characterizing N release patterns from EEFs. The method was reproducible and the result s were only slightly affected by variations in environmental factors such as microbial activity, soil moisture, temperature and texture. Information generated by the se studies sugges ts that by incubating the EEFs using the standard method at temperatures between 21 and 25C, reproducible and consistent N release curves can be obtained regardless of location. Short Term Accelerated Laboratory Extraction Method The objective of this study was to conduct method optimization and ruggedness testing of the accel erated lab extraction procedure by assessing the effect of various parameters at different levels on the performance of the method to estimate nutrient release rates of EEFs. The following hypotheses were tested in this study: 1) N, P and K release rates i ncrease with increasing extraction temperature and vice versa; 2) sample size affects N, P and K release rates, increasing and decreasing with the high and low fertilizer sample size respectively; 3) reducing the extraction time reduces the

PAGE 144

144 total amount of N, P and K extracted from the EEFs; and 4) small variations in several performance parameters of the method do not influence N, P, and K release rates of the EEFs. In general, varying the fertilizer sample size and reducing the extraction time did not aff ect N, P or K release rates from any EEFs. Temperature was the only factor found to substantially influence nutrient release rates from the EEFs studied. Raising the extraction temperature by 10 C at least doubled the N release rate at each extraction tim e. At the low extraction temperature of 50 C, less than 35% of the applied N was recovered in 168 hours, indicating that this temperature was too low to accelerate the N release profile of these EEFs. Differences in nutrient release rates between the mediu m and the standard temperature (55 vs. 60 C) depended on fertilizer coating thickness. The thickly coated fertilizers were more responsive to this minor temperature change (5 C), but they always provided lower N, P, and K release rates compared with the th inly coated fertilizers, suggesting that coating thickness highly influenced the amount of nutrient released. This finding was interesting because it showed that the accelerated lab extraction method has the capability to differentiate nutrient release pat terns from fertilizers having the same coating technology but different designed release rates Th e significant temperature effect was highly marked for P and to a lower extent for N and K. Total P released decreased to half when the maximum temperature de creased from 60 to 55 C. Based on these results, the optimal extraction temperature sequence that produced the highest N, P, and K release rates and showed no abnormal nutrient release due to coating deformation or fertilizer caking was determined to be:

PAGE 145

145 E xtraction #1 2 :00 hr. @ 25C ; Extraction #2 2 : 00 hr. @ 50C ; Extraction #3 20 :00 hr @ 55C ; and Extraction #4 50 :00 hr. @ 60C. Overall, the optimized method showed to be rugged for measuring N release rates of coated EEFs. Only t wo factors, flow rate of extract solution and extraction temperature, produced significant effects for N released from IBDU and RLCU (temperature only ), thus indicating that the method was little affected if slight variations from the standard conditions were applied. Neverthel ess, these factors should be carefully controlled to ensure high precision of the method. Validation of the extract flow rate and temperature at standard values is recommended before each extraction sequence. Ruggedness testing of this method for P and K w as inconclusive due to the use of an insufficient number of P and K based EEF materials. At least five representative test samples are recommended to be used in method performance studies (Horwitz, 1995). Therefore, further experiments are necessary to ful ly evaluate the effect of these factors on P and K released from coated EEFs. Correlation of the Soil Incubation and the Accelerated Extraction Methods The objective of this study was to develop a prediction model that can use data from the accelerated lab extraction method to predict the 180 d N release curve for a specific EEF product. It was hypothesized that the accelerated lab extraction method generates data that can be correlated with the N release data from the soil incubation method in order to pre dict the N release curve for a specific EEF. Based on the R 2 > 0.90 obtained for most EEF materials, results indicate d that the data generated from the accelerated lab extraction method c ould be used to predict the N release from these EEFs during a 180 d period under controlled, yet biologically active soil conditions. High R 2 indicated a strong relationship considering the limited

PAGE 146

146 amount of data used as well as the sampling and analytical variances. The accuracy of the final prediction model varie d by fe rtilizer type and technology but the two step r egr ession prediction method showed strong evidence that the accelerated laboratory extraction is a viable method to predict N release patterns of EEF with time including those that require biological activity for N release The predicted regression constants for ureaform did not produce as high of a coefficient of determination as the rest of the EEFs, but the regression accounted for 76% of the variation in the total sum of squares. This result was possibly d ue to the low N recovery from ureaform during 180 d (~30% of N applied). For any further statistical analyses, additional replications for each EEF are needed so that any potential bias can be identified and the quality of the estimates can be more precis ely quantified. These results are promising as no standard method to evaluate and regulate EEF materials has been accepted by AOAC. Based on these results, the accelerated laboratory method was capable of evalu ating and predicting the N release properties of the selected EEFs, thus, it ha s potential for use as a regulatory and evaluation method for a broad range of EEFs.

PAGE 147

147 LIST OF REFERENCES Ahmed, I.U., O.J. Attoe, L.E. Engelbert, and R.E. Corey. 1963. Factors affecting the rate of release from fertilizers from capsules. Agron. J. 55:495 499. Alexander, A. and H. Helm. 1990. Ureaform as a slow release fertilizer: A review. Z. Pflanzenernaehr Bodenkd 153:249 255. Allen, S.E. 1984. Slow release nitrogen fertilizers. In R.D. Hauck (Ed.), Nitrogen in crop produ ction ( pp. 195 206). ASA CSSA SSSA. Madison, WI. Allen, S. E., C. M. Hunt, and G. L. Terman. 1971. Nitrogen release from sulfur coated urea, as affected by coating weight, placement and temperature. Agron. J. 63(4):529 533. Alva, A. 1992. Differential lea ching of nutrients from soluble vs. controlled release fertilizers. Environ. Manage. 16(6):769 776. Alvares Ribeiro, L. and A. Machado. 1997. Usefulness of ruggedness in the validation of flow injection analysis systems. Analytica Chimica Acta 355:195 201. Al Zahrani, S. M. 2000. Utilization of polyethylene and paraffin waxes as controlled delivery systems for different fertilizers. Industrial & Engineering Chemistry Research 39(2):367 371. Anderson J P E 1982 Soil respi ration. In A.L. Page, R.H. Miller D.R. Keeney ( E ds ) Methods of soil analysis: chemical and microbiological properties, Part 2. Agronomy Monograph No. 9 ( pp 831 871 ). ASA SSSA Madison WI AOAC. 1965. Official methods of analysis, 10th edition. Association of Official Agricultural Chemi sts. Washington, D.C. APPFCO. 2004. Association of plant food control officials publication # 57. Association of Plant Food Control Officials, Inc. West Lafayette, IN. APPFCO 1995. In D.L. Terry (Ed.), A ssociation of plant food control officials publicati on # 48 (pp.164). Association of Plant Food Control Officials, Inc. West Lafayette IN Baligar, V.C., N.K. Fageria, and Z.L. He. 2001. Nutrient use efficiency in plants. Co mm. Soil Sci. Plant Analysis. 32 (7 &8 ): 921 9 50. Blouin, G.M., D.W. Rindt, and O.E. M oore. 1971. Sulfur coated fertilizers for controlled release: Pilot plant production. J. Agric. Food Chem. 19(5):801 808. Broschat, T.K., and K.K. Moore. 2007. Release rates of ammonium nitrogen, nitrate nitrogen, phosphorus, potassium, magnesium, iron, a nd manganese from seven controlled release fertilizers. Comm. Soil Sci. Plant Analysis. 38(7):843 850.

PAGE 148

148 Brown, M. J., R. E. Luebs, and P. F. Pratt. 1966. Effect of temperature and coating thickness on the release of urea from resin coated granules. Agron. J. 58(2):175 178. Cabrera, R.I. 1997. Comparative evaluation of nitrogen release patterns from controlled release fertilizers by nitrogen leaching analysis. HortScience 32(4):669 673. Canali, S., and A. Benedetti. 2006. Soil nitrogen mineralization. In J. Bloem, D.W. Hopkins, and A. Benedetti (Eds.), Microbiological methods for assessing soil quality (pp. 127 135). CABI Publishing. Cambridge, MA. Chen, D., J. Freney, I. Rochester, G. Constable, A. Mosier, and P. Chalk. 2008. Evaluation of a polyolefin coat ed urea (Meister) as a fertilizer for irrigated cotton. Nutrient Cycling in Agroecosystems 81(3):245 254. Christianson, C.B. 1988. Factors affecting N release of urea from reactive layer coated urea. Fert. Res. 16:273 284 Dai, J., X. Fan, J. Yu, F. Liu, an d Q. Zhang. 2008. Study on the rapid method to predict longevity of controlled release fertilizer coated by water soluble resin. Agricultural Sciences in China 7(9):1127 1132. Dejaegher, B., and Y. Vander Heyden. 2007. Ruggedness and robustness testing. J Chromatogr. A. 1158:138 157. Du, C., D. Tang, J. Zhou, H. Wang, and A. Shaviv. 2008. Prediction of nitrate release from polymer coated fertilizers using an artificial neural network model. Biosystems Eng. 99(4):478 486. Du, C., J. Zhou, and A. Shaviv. 2 006. Release characteristics of nutrients from polymer coated compound controlled release fertilizers. J. Polym. Environ. 14(3):223 230. Efron, B. The j ackknife, the b ootstrap and o ther r esampling p lans. 1 982. In CBMS NSF regional conference series in ap plied mathematics (pp. 21 26). The Society for Industrial and Applied Mathematics Philadelphia, P A Efron, B., and R.J. Tibshirani. 1993. An introduction to the bootstrap. In D.R. Cox, D.V. Hinkley, N. Reid, D.B. Rubin, and B.W. Silverman (Eds.), Monograp hs in statistics and applied probability 57 ( 31 37 ). Chapman & Hall /CRC. Boca Raton, FL Eickhout, B., A. Bouwman, and H. van Zeijts 2006. The role of nitrogen in world food production and environmental sustainability. Agriculture, Ecosystems & Environmen t, 116(1 2) : 4 14. FAO. 2008. Current world and fertilizer trends and outlook to 2011/12. Food and Agriculture Organization, United Nations. Rome, Italy. A vailable at: ftp://ftp.fao.org/agl/agll/docs/cwfto11.pdf

PAGE 149

149 Friedman, S. P., and Y. Mualem. 1994. Diffusi on of fertilizers from controlled release sources uniformly distributed in soil. Fert. Res. 39(1) : 19 30. Fujinuma, R., N.J. Balster, J.M. Norman. 2009. An improved model of nitrogen release for surface applied controlled release fertilizer. Soil Sci. Soc. Am. J. 73(6):2043 2050. Gambash, S., M. Kochba, and Y. Avnimelech 1990. Studies on slow release fertilizers: II. A method for evaluation of nutrient release rate from slow releasing fertilizers. Soil Sci.150(1):446 450. Gandeza, A. T., S. Shoji, and I. Ya mada. 1991. Simulation of crop response to polyolefin coated urea: I. Field dissolution. Soil. Sci. Soc. Am. J. 55(5):1462 1467. Giordano, P.M. and J.J. Mortvedt. 1970. Release of nitrogen from sulfur coated urea in flooded soil. Agron. J. 62:612 614. Goer tz, H. M. 1993. Controlled release technology. In M. How Grant (Ed.), Kirk Othmer encyclopedia of chemical technology 4th edition (251 274). John Wiley and Sons, Inc. New York City, NY. Hanafi, M. M., S. M. Eltaib, M. B. Ahmad, and S. R. Syed Omar. 2002. E valuation of controlled release compound fertilizers in soil. Comm. Soil Sci. Plant Analysis. 33(7):1139 1156 Hart, S.C., J.M. Stark, E.A. Davidson, and M.K. Firestone. 1994. Nitrogen mineralization, immobilization, and nitrification. In R.W. Weaver, S. Angle, P. Bottomley, D. Bezdicek,S. Smith, A. Tabatabai and A. Wollum (Eds.), M ethods of soil analysis, part 2: microbiological and chemical properties (pp. 985 1 016). Soil Science Society of America, Inc. Madison, WI. Horwitz, W. 1995. Protocol for the d esign, conduct and interpretation of method performance studies. Pure &Appl. Chem., 67(2):331 343. Huett, D.O. 1997a Fertilizer use efficiency by containerized nursery plants 1. Plant growth and nutrient uptake. Aust. J. Agric. Res. 48(2):251 258. Huett, D. O. 1997b Fertilizer use efficiency by containerized nursery plants: 2. Nutrient leaching. Aust. J. Agric. Res. 48(2):259 265. Huett, David O., and Beverley J. Gogel. 2000. Longevities and nitrogen, phosphorus, and potassium release patterns of p olymer coated controlled release fertilizers at 30C and 40C. Comm. Soil Sci. Plant Analysis 31(7):959 973. Hughes, T.D. 1976. Nitrogen release from isobutylidene diurea: soil pH and fertilizer pasticle size effects. Agron. J. 68:103 106.

PAGE 150

150 Husby, C.E., A. X. Niemiera, J.R. Harris, and R.D. Wright. 2003. Influence of diurnal temperature on nutrient release patterns of three polymer coated fertilizers. HortScience 38(3): 387 389. Jahns, T., H. Ewen, and H. Kaltwasser. 2003. Biodegradability of urea aldehyde c ondensation prod ucts. J. Polym. Environ. 11(4): 155 159. Jahns, T. and H. Kaltwasser. 2000. Mechanism of microbial degradation of slow release fertilizers. J. Polym. Environ. 8(1):11 16. Jarosiewicz, A., and M. Tomaszewska. 2003. Controlled release NPK fer tilizer encapsulated by polymeric membranes. J. Agric. Food Chem. 51(2): 413 417. Jarrell, W. M., and L. Boersma. 1979. Model for the release of urea by granules of sulfur coated urea applied to soi l. Soil Sci. Soc. Am. J. 43(5): 1044 1050. Joliffe, I.T. 200 2. Principal c omponent a nalysis. In Springer series in statistics (pp. 1 6). Springer Verlag New York Inc New York, NY. Kochba, M., S. Gambash, and Y. Avnimelech. 1990. Studies on slow release fertilizers:1.Effects of temperature, soil moisture, and wat er vapor pressure. Soil Sci. 149(6):339 343. Koivunen, M.E., and W.R. Horwath. 2004. Effect of management history and temperature on the mineralization of methylene urea in soil. Nutrient Cycling i n Agroecosystems 68(1): 25 35. Lamont, G.P., R.J. Worrall, and M.A. O'Connell. 1987. The effects of temperature and time on the solubility of resin coated controlled release fertilizers under laboratory and field conditions. Scientia Horticulturae 32( 3):265 273. Lubkowski, K., and B. Grzmil. 2007. Controlled rel ease fertilizers. Polish Journal of Chemical Technology 9(4):81 84. Lunt, O.R., and S.B. Clark. 1969. Properties and value of 1,1 diureido isobutane (IBDU) as a long lasting nitrogen fertilizer. J. Agric. Food Chem. 17(6):1269 1271. Lunt, O.R. and J.J. Oe rtli. 1962. Controlled release of fertilizer minerals by encapsulating membranes: II. Efficiency of recovery, influence of soil moisture, mode of application, and other considerations related to use. Soil Sci Soc. Amer. Proc. 26:584 587. McClellan, G.H. an d R.M. Scheib. 1973. Characterisation of sulphur coating on urea. Sulphur Inst. J. 9:3 4. Medina, L.C., T.A. Obreza, J.B. Sartain, R.E. Rouse. 2008. Nitrogen release patterns of a mixed controlled release fertilizer and its components. HorTechnology 18(3): 475 480.

PAGE 151

151 Nijhuis, A., H.C.M. van der Knaapa, S. de Jonga, and B.G.M. Vandeginste. 1999. Strategy for ruggedness tests in chromatographic method validation. Analytica Chimica Acta 391:187 202. Novillo, J., M.I. Rico, and J.M. Alvarez. 2001. Controlled relea se of manganese into water from coated experimental fertilizers: Laboratory characterization. J. Agric. Food Chem. 49(3):1298 1303. automated colorimetry. Environmental Monitoring Systems Laboratory, Office of Research and Development. U.S. Environmental Protection Agency, Cincinnati, OH, USA. Oertli, J.J. 1980. Controlled release fertilizers. Fert. Res. 1:103 123. Oertli, J.J., and O.R. Lunt. 1962. Controlled release of fertilizer minerals by encapsu lating membranes: I. Factors influencing the rate of release. Soil Sci. Soc. Amer. Proc. 26:579 583. Paramasivam, S., and A.K. Alva. 1997. Leaching of nitrogen forms from controlled release nitrogen fertilizers. Commun. Soil Sci. Plant Anal. 28(17&18):1663 1674. Patel, J., and G.C. Sharma. 1977. Nitrogen release characteristics of controlled release fertilizer during a four month soil incubation, J. Am. Soc. Hortic. Sci. 102(3):364 367. Prasad, M., and M. J. Woods. 1971. Release characteristics of nitrogen fertilizers in peat and san d. J. Agric. Food Chem 19(1): 96 98. Salman, O. A., G. Hovakeemian, and N. Khraishi. 1989. Polyethylene coated urea. 2. Urea release as affected by coating material, soil type and temperature. Industrial & Engine ering Chemistr y Research 28(5): 633 638. Sartain, J. B. and J.K. Kruse. 2001. Selected fertilizers used in turfgrass fertilization. Univ. of Fla. IFAS Soil and Water Sci. Dep. CIR 1262. Sartain, J.B., W.L. Hall, R.C. Littell, and E.W. Hopwood. 2004. New tools for the an alysis and characterizati on of slow released fertilizers In W.L Hall and W.P. Robarge (Eds.), Environmental impact of fertilizer on soil and water. Amer. Chem. Soc. Symp. Ser. 872 ( p p 180 195 ) Oxford University Press. Washington, DC. Savant, N.K., J.R. Clemmons and A.F. James. 1982. A technique for predicting urea release from coated urea in wetland soil. Comm Soil Sci. Plant Analysis 13:793 802. Shaviv, A., S. Raban, and E. Zaidel. 2003a. Modeling controlled nutrient release from polymer coated fertil izers: diffusion release from single granules. Environment al Science & Technology 37(10): 2251 2256.

PAGE 152

152 Shaviv, A., S. Raban, and E. Zaidel. 2003b. Modeling controlled nutrient release from a population of polymer coated fertilizers: statistically based model for diffusion release. Environmental Science & Technology 37(10):2257 2261. Shaviv, A. 2001. Advances in controlled release fertilizers. Advances in Agronomy 71:1 49. Simonne, E.H., and C.M. Hutchinson. 2005. Controlled release fertilizers for vegetable p roduction in the era of best management practices: teaching new tricks to an old dog. HortTechnology 15( 1):36 46. Soil Survey Staff. 2004. Particle size analysis method 3A1a1. In R. Burt (Ed.), Soil survey laboratory methods manual. Soil survey investigat ions report No. 42, version 4 (pp. 41). USDA NRCS National Soil Survey Center, Lincoln, NE. Stanford, G., and S. J. Smith. 1972. Nitrogen mineralization potentials of soils. Soil Sci. Soc. Am. J. 36(3):465 472. Stewart, W.M., D.W. Dibb, A.E. Jhonston, and T.J. Smyth. 2005. The contribution of commercial fertilizer nutrients to food production. Agron. J. 97 ( 1 ): 1 6 Tomaszewska, M. and A. Jarosiewicz. 2002. Use of polysulfone in controlled release NPK fertilizer formulations. J. Agric. Food Chem. 50(16):4634 4639. Trenkel, M.E. 1997. Controlled release and stabilized fertilizers in agriculture. International Fertilizer Industry Assn. (IFA). Paris, France. Trenkel, M.E. 2010. Slow and controlled release and stabilized fertilizers: An option for enhancing nutr ient use efficiency in agriculture. International Fertilizer Industry Assn. (IFA). Paris, France. Vander Heyden, Y., A. Nijhuis, J. Smeyers Verbeke, B.G.M. Vandeginste, and D.L. Massart. 2001. Guidance for robustness: ruggedness tests in method validation. J. Pharm. Biomed. Anal. 24:723 753. Wang, F.L. and A.K. Alva. 1996. Leaching of nitrogen from slow release urea sources in sandy soils. Soil Sci. Soc. Am. J. 60:1454 1458. Wilson, M. L., C. J. Rosen, and J.F. Moncrief. 2009. Potato response to a polymer c oated urea on an irrigated, coarse textured soil. Agron. J. 101(4):897 905. Yanai, J., T. Kosaki, A. Nakano, and K. Kyuma. 1997. Application effects of controlled availability fertilizer on dynamics of soil solution composition. Soil Sci. Soc. Am. J. 61(6):1781 1786. Youden, W.J., and E.H. Steiner. 1975. Statistical Manual of the Association of the AOAC, Association of Official Analytical Chemists. Arlington, VA.

PAGE 153

153 Zvomuya, F., C.J. Rosen, M.P. Russelle, and S.C. Gupta. 2003. Nitrate leaching and nitrogen recovery following application of polyolefin coated urea to potato. J Environ. Qual. 32(2):480 489.

PAGE 154

154 BIOGRAPHICAL SKETCH Carolina Medina was born in Guayaquil, Ecuador. She attended the Polytechnic School of the Littoral in Guayaquil, Ecuador, for two years and then transferred to the University of Florida where she received her ba a gricultural o perations m anagement in 2003. She continued further studies in the Soil and Water Science Department at the University of Florida, obtaining a m aster with specialization in s oil f ertility in May 2006. After earning her doctorate from the University of Florida in the spring of 2011 Carolina would like to work as a research scientist in the agriculture industry