Evaluation of Portable Soil Test Kits Promoted for Use by Smallholder Farmers to Make Site-Specific Fertilization Decisions
Recently, there is growing interest in updating fertilizer recommendations for crops in most developing countries, particularly in sub-Saharan Africa (SSA) and southern Asia. Many farmers in this part of the world have been correctly advised that investment in fertilizer inputs should be preceded by proper analysis of soil chemical and physical properties in order to ascertain which nutrients are suboptimal. Currently, the most frequently used approaches to soil analysis are those offered by standard soil analytical laboratories, mostly through wet chemistry, using various extraction and digestion methods. Given the constraints associated with standard soil analytical labs for smallholder farmers, a simpler and quicker approach is needed for soil tests at the farm level in the developing countries. Portable soil test kits with their mobile features and low costs offer attractive options capable of providing tailored and real-time fertilizer recommendations for smallholder farmers in remote regions and in areas without functional laboratories. Many soil test kits are available on the market in most parts of SSA and southern Asia. Their predictive capacity of soil fertility is not consistent, which suggests the need for proper testing, calibration, and validation of the kit outputs against reference wet chemistry laboratory data and response curve of the crop to fertilizer. The overall objective of this work was, therefore, to evaluate selected soil test kits in terms of their performance against standard soil analytical labs to ascertain the reliability of data generated with these kits. Specific objectives were to: (a) evaluate selected soil test kits with respect to their ability to accurately analyze the macronutrients (nitrogen [N], phosphorus [P], and potassium [K]) and pH; (b) compare them with respect to crop performance (dry matter and nutrient uptake); and (c) based on the results obtained, make recommendations for smallholder farmers, taking into account the above objectives. The following soil test kits were evaluated against standard laboratory tests: Hach, SoilDoc, and Kasetsart University soil test kits. The Kasetsart and Hach soil test kits were chosen based on their ease of use, performance, and cost. They have been used in several different projects in vii Rwanda, Ghana, Afghanistan, and Thailand and are currently being promoted by several projects in SSA and southern Asia. The third kit – SoilDoc – was added due to its widespread and aggressive introduction and use in SSA. Eight benchmark soils of different physicochemical properties were used to evaluate the soil test kits. During the second half of the FY16 project year, the soils were characterized by IFDC’s soil analytical lab and validated by Auburn University’s soil testing lab. Additionally, 40 archived soil samples from the North American Proficiency Testing (NAPT) Program (Soil Science Society of America quality control program for soil testing) were obtained from Auburn University. These samples were strategically selected to capture variations in soil texture, pH, and organic matter content and used to evaluate the accuracy of the portable soil test kits. Finally, a greenhouse experiment was set up to validate the soil test results coming from both the wet chemistry procedures of the soil analysis lab and those of the portable soil test kits by matching the soil nutrients, as determined by the soil test kits, to plant nutrient uptake. The study confirmed that the Kasetsart soil test kit was user friendly for most smallholder farmers. Proper use of the SoilDoc kit, on the other hand, would require the user to have considerable chemistry knowledge and experience in soil testing. The accompanying training manual did not provide some specific, detailed steps. This oversight could have led to erroneous results for people inexperienced in soil testing. However, it should be noted that SoilDoc advocates that its users participate in an intensive one-week training course prior to independent use. At the current stage of development, the SoilDoc kits should be handled by professional chemists familiar with soil testing or by personnel well-trained in using the kit. Results obtained from the analyses of the IFDC benchmark soils showed a good correlation (between the soil test kits and the standard wet chemistry analyses) for soil pH. Nitrate-N (NO3-N) showed reasonable correlation between standard chemistry and the soil test kit. However, in acidic soils, NO3-N values obtained with all three test kits were outside an acceptable range as determined by the KCl method for standard wet chemistry analysis. The results obtained for “available” P did not produce good correlations and were highly pHdependent. With soils of near-neutral pH, all three soil test kits produced P concentration values that were within an acceptable range of values for the standard lab analysis, but the kits performed poorly in acidic soils. The Hach soil test kit performed reasonably well for P with viii alkaline soil, whereas the other two performed poorly. Only the SoilDoc soil kit performed well for the determination of K. For the archived NAPT samples, the pH values determined using the three soil test kits were not significantly different from the pH values obtained by standard laboratory procedure. Of practical significance, the soil test kits more accurately differentiated among strongly acidic soils, acidic soils, near-neutral soils, and alkaline soils represented in the NAPT soil samples compared to IFDC’s benchmark soils. Although there was a very high correlation of the results of the nitrate analysis done with the SoilDoc soil test kit with those of the standard soil chemistry procedure (R2 > 0.96) for the NAPT samples, the actual values obtained from the SoilDoc soil test kits were consistently orders of magnitude greater than those obtained from the standard wet chemistry procedure. On the other hand, there was a poor correlation of the P concentration values with the standard laboratory Pi procedure. However, for the soil samples with near-neutral pH, there was a rather strong correlation of the “available” P values between the SoilDoc results and those of the standard laboratory procedure. Similarly to the soil nitrate determination with SoilDoc soil test procedure, the actual phosphorus values obtained were orders of magnitude (an average of six times) greater than those obtained from the standard wet chemistry procedure. For potassium determination, the SoilDoc was the only test kit that produced relatively good results, relative to the other two soil test kits. However, contrary to the results observed for nitrate and phosphorus, the actual potassium values obtained with the SoilDoc soil test kits were consistently smaller (about one-half) than those obtained from the standard wet chemistry procedure. The combined results suggest the need for a critical look into the extraction procedures for the various elements and the algorithms being used for the calculations within the SoilDoc software. As observed with the benchmark soils, the pH of the NAPT soils had a significant effect on the nitrate analysis using the Hach soil test kits. With the exception of the soils having strongly acidic pH, there was a very high correlation between the values obtained with the Hach soil test kits and those of the standard wet chemistry procedure. However, the actual soil nitrate concentration values from the Hach soil test kits were relatively smaller (an average of 65%) than those obtained with the standard wet chemistry procedure. For “available” P determination, ix the Hach soil test kit was the only one that produced good results. With the exception of the strongly acidic soil, determinations on most of the weakly acidic, near-neutral, and weakly alkaline soils produced values that were within the limits of those observed with the standard wet chemistry procedure. However, as observed with the benchmark soils for exchangeable K, the Hach soil test kit was not good enough to produce acceptable K values, compared with values obtained with the wet chemistry analysis. The results obtained with the Kasetsart soil test kit for the archived soil samples were rather inconsistent, as observed for the benchmark soils. For soils with medium to high nitrate concentrations, there was a partial match with the results of the standard wet chemistry procedure. However, there were several soil samples that were classified by the Kasetsart soil test kits as having very low to low nitrate concentrations that were contrary to the values obtained with the standard wet chemistry procedure. Similarly, for “available” P and potassium determinations, the kit produced good results for the soils with high concentrations but not for those with low to medium concentrations. There were samples with inherently low P and K concentrations that were designated as having high to very high concentrations when the Kasetsart soil test kits were used (and vice versa for some samples). This anomaly could be attributed to the soil pH and the organic matter content of the soil. The acidity of the soil likely affected the chemical extraction of the nutrients from the soil, but since the procedure used by the Kasetsart kit for analysis is entirely colorimetric, the organic matter content of the soil likely compromised the color of the soil extract, which affected the reading of the extract to determine elemental concentrations. Thus, the ranges of pH and organic matter content of soils within which the Kasetsart soil test kits produce accurate and acceptable values must be evaluated and specified in the user’s manual. Also, it is important to align the “available” P level designations by the Kasetsart soil test kits with the range of “available” P concentration values to be consistent with levels assigned by researchers for most soils. Thus, for now, soil fertility recommendations using portable soil test kits are not realistic. The value of these three kits is limited to giving baseline information on soil properties, particularly for pH, NO3-N, and “available” P for soils with a near-neutral pH. Also, by using the kits, farmers could, at least, identify the limiting and abundant nutrients within their field, assuming costs were minimal. Even if a consistently well-performing kit is eventually identified, before it x can be widely used, field validation should be required based on a soil test kit approach that includes the blanket recommended fertilizer as a control to which the soil test kit recommendation is compared in terms of yield-enhancing efficiency and economic profitability. The economic feasibility of such an approach is questionable, particularly in light of the recent focus on spectral analysis as a replacement for standard wet chemistry. Regardless, for the immediate future, laboratory wet chemistry will remain the standard for identifying soils’ fertility and for improving fertilizer recommendations.
Nutrient uptake, Soil testing