Articles
TESTING A NITROGEN ADVISORY MODEL FOR VEGETABLES
This is due to their high dry matter production over relatively short growing periods and to the poor ability of root systems to extract soil mineral nitrogen.
Nitrogen fertilization for best economic returns and optimal nutrient use efficiency should be adapted to local climate and local soil fertility in order to reduce losses to surrounding areas.
Monitoring of N-levels in plant and soil during the growing season is, however, expensive and/or impracticable.
Methods of calculating soil nitrogen availability and plant uptake would therefore be a useful alternative.
A model for predicting the optimum N-fertilizer level for a wide range of crops has been developed at Horticulture Research International in England (Greenwood and Draycott 1989). The validity of this model, and that of a more recent modified version, has been evaluated under Norwegian conditions, using data from multiharvest cabbage fertilization trials performed in 1993 and 1994 on soils with contrasting moisture-holding capacity.
Different basal N-fertilizer dressings were applied, ranging between 50 and 300 kg/ha N. Weekly model calculations were used to determine the timing of additional split applications which were required during the growing season in order to achieve predetermined levels of dry matter yield.
Model predictions of dry matter yield and nitrogen concentration agreed reasonably well with values measured during the growing season, although the original model seemed to overestimate plant nitrogen levels early in the growing season and underestimate them later on.
The more recent version of the model gave closer agreement with measured N-values.
Apparent recoveries of N-fertilizer were measured in 1994 by the inclusion of a zero treatment.
Values were in most cases close to 80 % which is the value assumed for cabbage by the model under optimal conditions.
Exceptions were found in some cases, for example when most of the nitrogen was applied at the start of the growing season.
Somewhat lower values were normally found on loamy soil than on sandy soil, possibly due to a difference in N-mineralization rate between these soils.
The use of the model enabled similar yield levels to be obtained with small basal dressings as with much larger ones, thus minimizing the risk of leaching during the growing season.
The model seems to be a useful tool for improving N-fertilizer recommendation for vegetable crops.
However, some of the model inputs still require better adaptation to Norwegian conditions.