Evaluation of extraction based fertilizer recommendations.

Abstract

There is a need to improve methods by which nitrogen, phosphorus and potassium are currently recommended. There is a considerable lack of mechanistical justifications for the methods used to recommend these nutrients. Lack of mechanistical justification can be attributed mainly to the disregard of nutrient (N, P and K) dynamics. Also the difficulty in incorporating these dynamics on fertilizer recommendation programs has compromised the mechanistical basis of extraction based approaches. The aim of the study was to evaluate these conventional (extraction based fertilizer recommendations) methods used to recommend these nutrients, by comparing their performance to the alternative approaches provided in this study. This evaluation was carried out through several studies, and a review of literature. From literature review it was apparent that there is indeed a need for revision of these methods. Their lack in mechanistical, technical and practical justification was considered and critically analyzed. It was proposed that alternative P and K recommendations can be achieved through quantity/intensity (Q/I) relations (amount of a respective nutrient in solution relative to the amount of nutrient adsorbed). It was also proposed that N recommendations can be improved by integrating mineralizable N. It was also concluded that these alternative approaches can routinely in a cost effective manner be determined. The first chapter evaluated P and K Q/I relations in several South African soils. Parameters of K dynamics were derived from activity ratio diagrams and these were used to explain K dynamics. Phosphorus sorption curves were linearized by Langmuir equation, and parameters derived therefrom were used to evaluate P dynamics. It was found that pH measured in water had a correlation coefficient (R2) of 0.71 with P sorption maxima. It was also found that electrical conductivity could account for 76% variance in K intensity parameter. It was suggested that these correlations could be exploited further to empirically model these crucial parameters. Thus, these correlations provide a possibility of determining these parameters routinely.

Pot trials were also conducted to evaluate the crop response, when P or K was made with the alternative approaches using maize and potato as test crops. Conventional extraction approach recommended higher P rates, and the P uptake between the two methods was not significantly improved. The extraction based approach recommended lesser K rates and K uptake was significantly higher under the alternative approach. The impact of integrating mineralizable N on N recommendations was also evaluated under control conditions. It was found that although alternative N recommendation approach recommended lesser N rates the N uptake was not significantly reduced. In fact the non-significant trend was that N uptake was higher when N recommendations were made with an alternative approach. From these initial pot trials only one nutrient was allowed to vary and the rest were kept constant at optimum levels. The second set of pot trials were carried out (parallel to the previous one), and on this set, all three nutrients were allowed to vary per experimental units. On these NPK was recommended with alternative approach and compared to the conventional approach. The results obtained were similar to those obtained when N, P or K were allowed to vary individually. It was also suggested that total carbon can be used to assess the validity of these approaches. This was based on the consistent inverse correlation that was obtained between total carbon and P or K. Field trials were also conducted at Ukulinga research farm Pietermaritzburg and Wartburg, using maize and potato as test crops. The lack of concurrent response from nutrient uptake was also observed here similar to the observations already made in pot trials. These were characterized by conventional method recommending higher rates of N and N uptake not concurrent with the rates. It was also found that there was a poor correlation between applied fertilizer and extraction based intensity parameters, with R2 ranging between 0.005 – 0.011, compared to R2 of Q/I parameter which was 0.98 for both P and K. This poor correlation was evident between nutrient uptake and total biomass. Yield of both maize and potato at both sites was higher when recommendations were made by alternative approaches, and yield grade of potatoes was also improved when the recommendations were made by alternative approach. Total biomass of maize was also significantly improved when the recommendations were made by the alternative approach. Earlier,

observation with regards to correlation of total carbon and nutrients was also observed under field conditions. This suggested that this is an important parameter to evaluate fertilizer recommendation program. It was concluded that recommending P and K with Q/I relations, and integrating mineralizable N on N recommendations is more mechanistically, technically, theoretically and practically justified compared to the conventional method.