Options to reduce ranges in critical soil nutrient levels used in fertilizer recommendations by accounting for site conditions and methodology: A review

Fertilizer recommendations (FR) to improve yields and increase profitability are based on relationships between crop yields and soil nutrient levels measured via soil extraction methods. Within these FR, critical soil nutrient (CSN) levels are used to distinguish nutrient deficient from non-deficient soils. The variation in CSN levels is large, implying a risk of over- or under-fertilization. Here, we review and assess the factors influencing the derivation of CSN levels in order to increase both their reliability and applicability within FR systems. The evaluated factors included site conditions, i.e., crop type and location as a surrogate for climate and soil properties, and methodological factors, i.e., the experimental approach (field or pot experiments), and statistical methods and cut-off point. Results showed that the range of values used to define the medium soil fertility classes coincided with the range of CSN levels derived from experimental data. We show that harmonizing methodological aspects can substantially reduce the uncertainty in the CSN levels (> 50%), implying a substantial enhancement of the reliability of FR systems. Inclusion of site conditions might further improve the reliability. To enable reduction in CSN levels requires well-documented field experiments and standardization of data collection and analysis. We foresee the potential for generic FR systems that make use of reliable data, more process-based interpretation of nutrient pools and accounting for the interactions among nutrients.