NPK fertilizer management for corn: Decision aids and test kits

Attanandana,T1., C. Suwannarat1, T. Vearasilp2, S. Kongton2, R. Meesawat3, P. Bunampol3, K. Soitong4 , C. Tipanuka4 and R.S. Yost5
1 2

Department of Soil Science, Kasetsart University, Bangkok, Thailand Land Development Department, Bangkok, Thailand 3 Department of Agriculture, Bangkok, Thailand 4 Department of Agricultural Extension, Bangkok, Thailand 5 Department of Tropical Plant and Soil Sciences, University of Hawaii, U.S.A. Abstract A simple and rapid soil test for NPK is being developed for use in the field. Correlation studies were undertaken to select a single extracting solution for NPK determination in soils for routine analysis. A highly significant correlation between Mehlich 1 extractable ammonium+nitrate and corn dry weight (CDW) was obtained (r values of 0.74**). The highly correlation between ammonium+nitrate with N uptake with r values of 0.77** was also observed. The highest, significant correlations were between extractable K and both CDW and K uptake using Mehlich 1. Extractable P and CDW and extractable P and P uptake were also significantly correlated. Mehlich 1 was selected for the extraction of the three nutrient elements (N, P and K). The colorimetric determination of ammonium, nitrate, and phosphorus by spectrophotometer were modified and developed for use with a standard color chart. Suitable correlations were obtained between spectrophotometer and color chart for the NPK determination. The process of soil analysis was simplified and developed to be soil test kit. In the field experments, the actual and predicted yield of corn (Zea mays, L.) in seven soil series across four provinces was compared. The N fertilizer quantity was generated by the CERES-Maize model while P and K were calculated by a MitscherlichBray equation. An agreement index of 0.86 between actual and predicted yield was obtained, indicating excellent correspondence. The PDSS (Phosphorus Decision Support System) was used to predict the P requirement and was tested with the 15 plots of the farmers field in the 5 provinces of corn production area and the four on-farm experiments in Nakorn Ratchasima province. The amount of fertilizer P predicted by PDSS was onehalf of the amount obtained by calculation of Mitscherlich-Bray equation. The yield of the CERES-Maize-PDSS prediction was higher than with the farmers practice. The economic analysis indicated that the CERES-MB and CERES-PDSS treatments gave the higher benefit compared to the farmers practice. Introduction Soil analysis is a widely adopted method for the evaluation of soil fertility and as a tool for fertilizer recommendations. In developing countries, technology transfer is restricted due to a lack of equipment, buildings, the specialized training and time consuming supportive research. Moreover, soil testing is also costly, poor farmers can not usually afford to have the test for initial nutrient content in all their fields. Soil testing for N, P, K analysis requires nutrient specific chemical extractants, determination methods, expensive instrumentation and training. Universal extractants for P, K, Ca, Mg, and other micronutrients have been proposed and widely used in the United States. Inductively Coupled Plasma is used for determinations for most of the extractable nutrients (Jones,

2
1985). However, a simple, rapid and accurate method of soil testing which is applicable to be used in the field has not been developed in Thailand. The purpose of this research was to search for a single extracting solution for NPK in soils and to evaluate the determination of soil ammonium, nitrate and potassium. Stage 1 of the investigation was to compare the effectiveness of various extraction methods for P to extract NPK in some important soils for growing corn in Thailand, e.g. Bray 2, citric acid, Mehlich1, Morgan, Olsen, and sodium lactate (Kamprath and Watson, 1980; Thompson and Pratt, 1954; Page et.al. 1982). Stage 2 was to develop the soil test kit for NPK. Stage 3 was to develop the NPK fertilizer recommendation system using modeling program and employing soil test kit for initial nutrient evaluation. Materials and Methods Pot experiment The experiment was 15 x 5 factorial in completely randomized design with 3 replications. Five soil series which are representative of soils of the major corn production area of Thailand were used in this study. They are Chai Badan series (Cd), fine, smectitic, isohyperthermic Leptic Haplusterts, Takhli series (TK), loamy-skeletal, carbonatic, isohyperthermic Entic Haplusterts, Lop Buri series (Lb), very fine, smectitic, isohyperthermic Typic Haplusterts, Satuk series (Suk), fine, loamy, siliceous, isohyperthermic Typic(Kandic) Paleustults and Pak Chong series (Pc), very fine, kaolinitic, isohyperthermic Rhodic Kandiustox. Some chemical properties of these soils are shown in Table 1. Three composite samples of each soil, which differed in P content, were collected. The fifteen soil samples were used for greenhouse and laboratory study. The treatments were control, (PK), (NK), (NP), and NPK. Suwan 5 corn variety was used as the test crop. Dry weight of corn and NPK uptake in the plants were recorded. Table 1 Physical and chemical properties of soils used in the study Soil Series Cd H Cd M Cd L Tk H Tk M Tk L Lb H Lb M Lb L Suk H Suk M Suk L pH1 8.1 7.9 8.2 8.2 8.1 8.1 8.2 8.0 8.1 4.9 5.2 5.4 Texture2 SCL C CL SCL C C C C C LS SCL LS Organic3 matter (g kg-1) 22 30 22 44 34 41 33 22 33 04 11 05 P4 K5 Ca5 Mg5 -1 -1 -1 (mg kg ) (mg kg ) (mg kg ) (mg kg-1) 70 28 11 81 44 20 60 27 10 15 5 3 280 140 80 200 130 210 280 90 90 40 90 40 11000 10000 13000 12000 18000 17000 13000 15000 16000 160 520 320 340 260 200 210 220 260 350 520 530 42 170 80

Pc H 7.2 C Pc M 5.7 C Pc L 6.4 C 1 1:1 soil:water ratio 4 Bray II extraction method

35 48 110 3000 160 32 22 420 3000 220 18 9 90 3600 220 2 3 hydrometer method Walkley-Black method 5 1 M NH4OAc extraction method

3
Extractant selection The soils were extracted using ten different extracting solutions and associated methodologies (Table 2). In each case, the resultant filtrate was analyzed for NH4, NO3, P and K by conventional methods. The N content in the plant was analyzed by the Kjeldahl method, and P and K by double acid digestion (Jones et. al. 1991). The uptake of NPK in the corn plants was calculated and the results were correlated with extractable NH4, NO3, P and K in the soils. This was repeated for each extraction method. Each extraction method was assessed by correlating extracted nutrient values with dry matter weight and nutrient uptake of corn in pot experiments. The single extraction and rapid determination methods will be further developed into a soil test kit that can be used in provinces where soil testing laboratories are not available. Table 2 The extracting solutions used in the study Method Bray 2 Citric acid Mehlich 1 Modified Mehlich 1 Morgan Olsen Ammonium lactate Sodium lactate 1
Sodium lactate 2

Extracting solutions 0.03 M NH4F + 0.1 M HCl, 1:10 soil : solution ratio, shake for 1 minute. 1 % citric acid, 1:20 soil : solution ratio, shake for 30 minutes. 0.05 M HCl + 0.0125 M H2SO4, 1:5 soil : solution ratio, shake for 5 minutes. 0.05 M HCl + 0.125 M H2SO4, 1:5 soil : solution ratio, shake for 5 minutes. 0.54 M NH4OAc + 0.7 M NaOAc pH 4.8, 1:10 ] soil : solution ratio, shake for 30 minutes. 0.5 M NaHCO3, pH 8.5, 1:20 soil:solution ratio, shake for 30 minutes. 0.111 M lactic acid + dil. acetic acid + dil. NH4OH, 1:20 soil:solution ratio, shake for 4 hrs. 0.111 M lactic acid + dil. acetic acid + dil. NaOH 1:20 soil : solution ratio, shake for 30 minutes. 0.111 M lactic acid + dil. acetic acid + dil NaOH 1.20 soil : solution ratio, shake for 4 hrs. 1 M NH4HCO3+0.005M DTPA(pH 7.6), 1:2 soil:solution ratio, shake for 15 minutes

Reference Kamprath and Watson, 1980 Thompson and Pratt, 1954 Jones, 1985 Modified from Jones, 1985 Kamprath and Watson, 1980 Page et. al. 1982 Riehm, 1959 Modified from Riehm, 1959 Modified from Riehm, 1959 Jones, 1985

NH4HCO3+DTPA

4
Process of DSSAT-CERES-Maize for N fertilizer recommendation The data of 10 soil series of Petchaboon, Lop Buri, Nakorn Sawan and Nakorn Ratchasima provinces were updated and the data of 28 soil series was taken from the database of Land Development Department. Climatic database including solar radiation, maximum and minimum temperature, rainfall intensity, frequency and annual distribution were obtained from the Meteorology Department. With the use of the Weatherman program, long term climatic data were used to predict climatic characteristics for the 19972001 period of study. A genetic coefficient study was performed using Suwan 5 and Suwan 3601 corn varieties as the test crops. All necessary data was recorded and calculated to estimate the genetic coefficients of the two corn varieties. Process of PDSS program for P requirement The intended crop, % clay content, and soil test P of each soil are the only inputs into the program. The program will generate the P requirement for a typical yield. With additional inputs of fertilizer cost, grain price, and interest, estimates of benefit/cost are calculated. NPK fertilizer recommendation development The fertilizer recommendation system was developed using existing decision aids. The DSSAT-CERES-Maize program version 3.0 was used for N fertilizer recommendation. Phosphorus requirement was predicted using PDSS modeling. The on-farm test using soil test kit and NPK fertilizer recommendation system The farmers field was selected to conduct the on-farm test. There are four sites, 1.6, 3.8, 0.8 and 2.1 ha. The soils were Lam Phaya Klang(Lg), Chatturat(Ct), and Lop Buri(Lb) series respectively. Table 3 showed the pH, texture, soil series and area of the four sites. Table 3 pH, texture, soil series and area the four sites, on- farm test Farmer Saweang Thonglang Oui Perm Soil series Lg Ct Ct Lb pH 7.5 7.0 7.0 8.0 Texture C L L C Area(ha) 1.6 3.8 0.8 2.1

Results and Discussion Extractant Selection The Mehlich 1 extracting solution gave correlation coefficients of 0.74**, 0.50* and 0.66** for NH4 plus NO3 and CDW, P and CDW and K and CDW respectively. Similarly, correlation coefficients of 0.55*, 0.59* and 0.64** were obtained for NH4 plus NO3 and CDW, P and CDW and K and CDW respectively when using the Morgan extracting solution. In turn, sodium lactate 1 extracting solution was associated with correlation coefficients of 0.77**, 0.71** and 0.60* for NH4 plus NO3 and CDW, P and CDW and K and CDW respectively (Table 4). Similar trends were obtained when NH4 plus NO3, P and K soil extracted values were correlated with corn uptake of these nutrients. This is with the exception of extractable K and K uptake as determined by sodium lactate 1, which resulted in a correlation coefficient of 0.48ns (Table 5). The correlation results indicate that the Mehlich 1, Morgan and sodium lactate 1 extracting

5
solutions are the most promising for NH4, NO3, P and K extraction. Mehlich1 was chosen as the single extracting solution due to highly significant correlation between extractable NH4 +NO3, extractable K and CDW, N and K uptake. Correlation coefficients of 0.74** and 0.77** were obtained for NH4 plus NO3 and CDW and N uptake, respectively. Correlation coefficients of 0.66** and 0.93** were obtained between extractable K and both CDW and K uptake. A significant correlation between extractable P and both CDW and P uptake was obtained. A correlation coefficient of 0.50* and 0.56* was obtained between extractable P and both CDW and P uptake. Table 4 Correlation coefficients between NH4 + NO3, P and K extracted by 10 different methods and dry weight of corn Method Bray 2 Citric acid Mehlich 1 Modified Mehlich 1 Morgan Olsen Ammonium lactate Sodium lactate 1 Sodium lactate 2 NH4HCO3 + DTPA Dry weight of corn (28 days after planting) NH+4 + NO-3 P 0.64** 0.22 0.50* 0.17 0.59* 0.44 0.58* 0.71** 0.65** 0.49 K 0.62** 0.65** 0.66** 0.63** 0.64** 0.63** 0.46 0.60* 0.33 0.62** 0.18 0.74** 0.21 0.55* 0.77** 0.10 -

Remarks: * significant correlation at 95% level ** significant correlation at 99 % level Table 5 Correlation coefficient between NH4 + NO3, P and K extracted by 10 different methods and nutrient uptake of corn Method Bray 2 Citric acid Mehlich 1 Modified Mehlich 1 Morgan Olsen Ammonium lactate Sodium lactate 1 Sodium lactate 2 NH4HCO3 + DTPA Nutrient uptake NH+4 + NO 3 0.17 0.77** 0.28 0.52* 0.81** 0.19 -

P 0.69** 0.21 0.56* 0.13 0.71** 0.51* 0.73** 0.82** 0.75** 0.63**

K 0.96** 0.96** 0.93** 0.95** 0.95** 0.89** 0.42 0.48 0.41 0.93**

Remarks: * significant correlation at 95% level ** significant correlation at 99 % level

6
Soil test kit development The colorimetric determination of ammonium, nitrate, and phosphorus by spectrophotometer were modified and developed for use with a standard color chart. Suitable correlations were obtained between spectrophotometer and color chart for the ammonium determination. For example, on 244 acid soils the correlation coefficients were 0.54**, 0.87** and 0.96** for clayey, loamy and sandy soils. On 41 alkaline soils the correlation was similarly high 0.79 **and 0.94 ** on clayey and sandy soils. Similar results were obtained in the case of nitrate and phosphorus (Table 6). In the case of potassium determinations, a highly significant correlation of NH4OAc extractable K determined by A.A. was obtained with the amount extracted by Mehlich1 and determined by colorimetric methods (Table 7). The process of soil analysis was also simplified, e.g. the soil was scooped instead of weighed. All necessary equipment was replaced by simple plastic bottles and droppers. Table 6 Correlation coefficients for Mehlich 1 extractable ammonium, nitrate, and phosphorus between the spectrophotometer and color chart using 285 soil samples Color chart Spectrophotometer ammonium nitrate phosphorus sandy r = 0.96** n = 51 r = 0.99** n = 51 r = 0.77** n = 51 acid soil loamy r = 0.87** n = 55 r = 0.98** n = 55 r = 0.54** n = 55 clayey r= 0.54** n= 138 r= 0.80** n = 138 r = 0.48** n = 138 alkaline soil sandy r = 0.94** n=8 r = 0.93** n=8 r = 0.83** n=8 clayey r = 0.79** n = 33 r = 0.86** n = 33 r = 0.50** n = 33

Table 7 Correlation coefficients between K extractable by Mehlich 1 (Colorimetric) and NH4OAc (A.A) using 285 soil samples Color chart A.A. potassium sandy r=0.75** n=51 acid soil loamy r=0.60** n=55 clayey r=0.73** n=138 alkaline soil sandy clayey r=0.67** r=0.89** n=8 n=33

The accuracy of the test kit The readings of soil nitrate, phosphorus and potassium of 15 samples by spectrophotometer and test kit were compared. The results showed that the test kit and spectrophotometer give the interpretation (Low, Medium, and High) in 14 of 15 soils, 13 of 15 soils and 13 of 15 soils for nitrate, phosphorus and potassium compared with the spectrophotometer reading (Table 8).

7
Table 8 Soil test data of the Pioneer Company's plot interpretations. No series NO3- content Spectrophotometer Test Kit Inter mg kg-1 pretation N 2.00 VL VL 18.00 L L 3.47 VL VL 4.38 VL L 4.37 VL VL 2.67 VL L 12.92 L L 7.00 VL L 3.00 VL VL 18.00 L L 1.25 VL VL 1.56 VL VL 15.00 L M* 12.00 L L 12.00 L L 14/15 (before planting) and their

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Lb Lb Lb Ln Ln Tk Tk Pc Ct Lb Cu Lb Wi Tk Pc

P content K content Spectrophotometer Test Test Kit A.A Kit mg kg-1 mg kg-1 Inter mg kg-1 pretation P K K 4.50 M H* M 80 0.25 VL VL H 130 3.50 M H* M 82 6.75 M M M 89 1.00 L L M 71 3.25 L L H 277 0.56 L VL H 174 6.00 M M L 39 2.00 L L M* 266 19.60 VH H H 628 10.00 VH H L 69 47.50 VH VH L* 84 4.41 M M H 106 1.25 L VL H 126 9.00 H H M 78 13/15 13/15

NPK fertilizer recommendation using the CERES-Maize model The predicted and measured yield of Suwan 3601 hybrid corn on some important soils in the four provinces of the corn belt area using the NPK fertilizer recommendations are shown in Table 9. An agreement index of 0.86 indicated the close agreement between the predicted and actual yield for the seven series (Willmott, 1982). The N fertilizer recommendation was determined by the CERES-Maize algorithm while the P and K fertilizer recommendations were from Mitscherlich-Bray equation*. (Dept. of Agriculture, 1966, 1967). In the process of making fertilizer recommendations using decision-aids, the Table 9 Predicted and actual yield of Suwan 3601, tested in the field and the agreement i n d e x v a l u e
Soil series Cd Tw Tk Wi Wi Sat Suk Ct Province Nakorn Sawan Lop Buri Lop Buri Lop Buri Petchaboon Petchaboon Nakorn Ratchasima Nakorn Ratchasima Actual yield 7,225 6,144 5,225 6,469 6,413 6,031 5,900 7,481 Predicted yield kg/ha 6,563 6,225 5,475 6,181 6,238 6,194 6,513 7,000 N-P2O5-K2O 94-75-0 94-75-31 125-0-31 125-31-63 125-75-63 94-106-63 156-75-63 94-0-31

Remarks : *log (100 - y) = log 100 - 0.05419 b - 0.03864 x - P requirement log (100 - y) = log 100 - 0.00618 b - 0.05132 x - K requirement y = relative yield, b = soil test value, x = fertilizer requirement

8
soil was identified for soil series and a composite sample was tested for nitrate, phosphorus and potassium content. The NPK fertilizer recommendation was then prepared according to the initial nutrient contents and soil series in each province. NPK fertilizer recommendations using CERES-Maize and PDSS models The fertilizer recommendation system was further developed by introducing the PDSS program for P requirement. For this on-farm study, the CERES-Maize was used for N fertilizer recommendations, the K fertilizer recommendation was derived from a Mitscherlich-Bray equation and the P recommendation was developed using the PDSS system. The four on-farm tests were performed comparing: 1) The farmers practice, 2) NPK fertilizer recommended by CERES-Maize-Mitscherlich-Bray and 3) NPK fertilizer recommended by CERES-Maize, PDSS and Mitscherlich-Bray. Table 10, and 11 show the initial nutrient level, fertilizer recommendations, yields of corn, and P after harvest of the on-farm test. The results indicated a higher yield of corn where the NPK fertilizer recommendation was developed using the decision-aids except on one site that was affected by the stem borer, which resulted in a yield that was low compared with the farmers practice. The phosphorus content in the soils after harvest of the CERES-MaizePDSS treatment was medium to high resulting from the addition of the recommended amount of fertilizer. In the case of the farmers field, one site, however, indicated a low content of P after harvest. Phosphorus fertilizer recommendations by the MitscherlichBray equation were about twice the amounts of P fertilizer recommended by the PDSS decision-aid. Table 10 Initial nutrient level, fertilizer recommendation, yield of corn and P after harvest of the farmers practice and CERES-Maize-PDSS treatments
Series Lg Ct Ct Lb pH 7.5 7.0 7.0 8.0 texture C L L C nutrient level
VL-VL-H VL-VL-H VL-VL-H VL-VL-H

Farmers practice Fertilizer yield P after applied kg/ha harvest 25-25-0 2769 L 13-19-0 4569 H 19-25-0 2925 H 69-38-0 2706 H

CERES-Maize PDSS Fertilizer yield P after applied kg/ha harvest 94-44-0 6063 M 94-50-0 4138 H 94-50-0 4469 H 125-69-0 3425 H

Table 11 Initial nutrient level, fertilizer recommendation, yield of corn and P after harvest of the CERES-Maize-Mitscherlich Bray treatment Farmer Sawean g Thong Oui Perm Series Lg Ct Ct Lb pH 7.5 7.0 7.0 8.0 texture C L L C Nutrient Level VL-VL-H VL-VL-H VL-VL-H VL-VL-H CERES-Maize Mitscherlich-Bray Fertilizer yield P after applied kg/ha harvest 94-94-0 5913 M 94-94-0 94-94-0 125-94-0 3650 4944 4394 H H H

Economic analysis The profit of the four on-farm tests calculated as the benefit (crop price x the yield increase minus the fertilizer cost, including basic cost). It is clearly seen that the CERESMaize-Mitscherlich-Bray (CERES-MB) and CERES-PDSS treatments gave higher profit

9
compared to the farmers practice (Table 12). Probably the profit would have been greater, however, one of the plots was strongly attacked by stem borer and the yield was quite low. Table 12 Economic analysis of the four on-farm test Oui Perm Average $US/ha Farmer's practice 0.8 180.1 17.6 -25.4 43.3 * CERES-MB 244.8 27.9 151.9 89.8 128.6 CERES-PDSS 285.5 97.9* 129.6 10.2 130.8 *The profit was low due to the corn plants were attacked by the stem borer **1US$=43.0 baht Conclusion Mehlich 1 extracting solution is proposed as a single extracting solution for NH4, NO3, P and K in soils of Thailand. Standard color charts for ammonium, nitrate, phosphorus and potassium were developed and used with a soil test kit for NPK in the field. Nutrient interpretations obtained by the soil test kit were comparable to values obtained in the laboratory. The results showed an agreement of 14/15, 13/15 and 13/15 for nitrate, phosphorus and potassium between soil test kit and spectrophotometer interpretations. The predicted and actual yield of NPK fertilizer recommendation using CERES-Maize program and Mitscherlich-Bray equation was performed in the field and the agreement index of 0.86 was obtained. In the case of P requirement by PDSS showed that P fertilizer can be reduced about one half the amount obtained from calculation of Mitscherlich-Bray equation. Using fertilizer recommendation with decision aids and test kits, higher profit was obtained compared to the amount in which the farmers use which have not taken the initial amount of nutrients and the sustainability of the soils into account. Recommendation for the further study The on-farm test has to be further expanded for larger area of corn production. The adequate amount of nutrient applied to the soil for corn production and at the same time enough for sustainable agriculture should be investigated. Moreover, economic analysis should be focussed and emphasized. The simplified program for NPK fertilizer recommendation should be loaded in the handheld computer for the extension worker as the aids for agricultural advisor to the farmers. Acknowledgement The authors would like to express their sincere thanks to the Thailand Research Fund (TRF) for the financial support of this study. References Department of Agriculture. 1966. Annual Report 1966. Bangkok, Thailand. Department of Agriculture. 1967. Annual Report 1967. Bangkok, Thailand. Jones, Jr. J. B. 1985. Soil testing and plant analysis: Guide to the fertilization of horticultural crops. Horticulture reviews. Vol. 7. A.V. Publishing Co. 68 p. Jones, Jr. J. B., B. Wolf and H.A. Mills. 1991. Plant analysis handbook, a practical sampling, preparation, analysis, and interpretation guide. Micro-Macro Publishing, Inc. 213 p. Sawaeng Thonglang

10
Kamprath, E.J. and M.E. Watson. 1980. Conventional soil and tissue tests for assessing the phosphorus status of soils. The role of phosphorus in agriculture. Edited by Khasawneh, F. E., E. C. Sample, and E. J. Kamprath. 1980. ASA-CSSA-SSSA, 677 South Segoe Road, Madison, WI 53711, USA. 910 p. Page, A. L., R. H. Miller and D.R. Keeney. 1982. Methods of soil analysis. part 2 Chemical and microbiological properties. Second edition. American. Society of Agronomy, Inc., Soil Science Society of America, Inc., Publisher, Madison, Wisconsin, USA. 1159 p. Riehm, H. 1959. Die Ammoniumlaktat-Essigsaure-Methode zur Bestimmung der leicht loslichen Phosphorsaure in karbonathaltigen Boden. Agrochemica 3: 49-65 Thompson, L.F. and P. F. Pratt. 1954. Solubility of phosphorus in chemical extractants as indexes to available phosphorus in Ohio soils. Soil Sci. Soc. Amer. Proc. 18(4): 467-470 Willmott, C.J. 1982. Some comments on the evaluation of model performance. American Meteorological Society 63:1309-1313. ----------------------