Evaluating a crop nitrogen simulation model, N_ABLE, using a field experiment with lettuce

Abstract

A field experiment with lettuce was carried out to evaluate the simulation model, N_ABLE, which has been widely used to predict soil mineral nitrogen requirements and potential leaching hazards for vegetable and arable crops in England and parts of Western Europe. Plant and soil were sampled regularly and dry weight (W), percent N in dry matter and soil mineral N (soil-N) were measured. Measured W and soil-N were compared with data simulated using N_ABLE both during growth and at final harvest. Dry weight followed an asymmetrical S-shaped curve when the growth period was either 57 or 61 days for all N levels. This implies that N_ABLE, which assumes a J-shaped growth curve, can only be used in the first three-quarters of the growing period. Simulated soil-N in the 0–30 cm layer corresponded well with measured values throughout the experiment when parameters for the recovery of soil mineral N (REC) and mineralisation rate of soil organic-N (NR) were set at 0.70 (i.e. 70%) and 0.86 kg ha-1 d-1 respectively, both calculated from field data, and were higher than default values. For longer periods of growth, the best fit was obtained using a modified asymmetrical S-shaped growth curve equation dW/dT = k2W Gf Gk /(1+ W), where k is a growth rate coefficient, G (≤ 1) is a correction coefficient to allow for any restriction in growth rate caused by sub-optimal%N in the crop and G= ( W/W)n is another correction coefficient to adjust the growth rate which is decreased caused by genetic or other reasons in the later part of the growth period. The S-shaped equation was examined by a lack of fit test, and the results showed that the residual errors ( SS = ∑( y-x)2, where x =simulated values, y = measured values) were not significantly different from experimental error, indicating that the S-shaped equation gave a good description of growth for the different N levels through the growth periods.