This thesis focuses on the optimisation of N fertiliser application, taking into account spatially variable soil conditions. Spatial soil variability effects both cropproduction and nitrate leaching. Site specific management tries to address spatially variable conditions. Research on site specific management was done for a field on the experimental farm "the Van Bemmelenhoeve" in the north westernpart of the Netherlands.
In summary the objectives of this study were:
1. To measure spatial and temporal variability of soil physical properties, yield and N for a farm field.
2. To quantify spatial and temporal variability, using simulation modelling and geostatistics, of:
(a) crop production and nitrate leaching.
(b) effects of N fertiliser management on production and leaching.
3. To identify areas in the field with comparable behaviour with respect to production and leaching, considering the effects of multi-year weather variation in both the growing season and the season with precipitation surplus.
4. To define representative profiles for the identified map units using two land quality indicators: nitrate leaching and crop production.
Chapter 2 serves as an introduction to the study area. The spatial variation of potato production, total N and soil physical properties are quantified. Potato yields, measured in 65 small plots varied between 30 and 45 tons ha -1, while yields of commercially attractive large potatoes varied between 3 and 15 tons ha -1. Such differences are economically significant for a farmer. Total N in the early part of the growing season varied between 21 and 53 kg ha -1. When compared with recommended fertiliser rates obtained from one mixed sample for the entire field, local over- and underfertilization can be demonstrated. These are bound to lead to groundwater pollution and inefficient production. Modeling can be used to balance production and environmental aspects in a quantitative manner. Exploring the spatial and temporal behaviour of the field using simulation models and geostatistics is shown to be an attractive approach in agrosystems research.
In chapter 3 the implications on farm management of spatially variable soil conditions are discussed. Maps displaying relevant soil variability are needed to guide site specific practices. In section 3.1 simulated water-limited yield patterns are shown to be stable in space and time. A "prototype" pattern is derived from seven simulated water-limited yield patterns. This "prototype" can be used as a base map for site specific management. The effects of site specific fertiliser application on potato production and nitrate leaching are studied using dynamic simulation in section 3.2. Simulations show that the recommended fertiliser dose for 1994 was too high for the entire field. Fine-tuning of the fertiliser application using the "prototype" pattern was succesful in maintaining high production levels and reducing leaching of nitrate to the groundwater. Leaching of nitrate occurs mainly during the "wet" winter season (September 15 - April 15). The amount of nitrate leached during the "wet" period is dictated by postharvest soil N conditions. In section 3.3 the soil-N contents are defined for 65 soil profiles corresponding with a predefined threshold value for nitrate leaching to the ground water. Leaching was calculated for 5 different N profiles at the beginning of the wet period for 20 years. For the given initial N range (15 - 120 kg N ha -1m -1) space time relations for the defined period are linear. The required N profile at September 15 is calculated using three risk levels of exceeding these thresholds. Spatial interpolation of the required N profiles results in N target maps which can be used to focus and evaluate N fertiliser management.
Farm management should focus N fertiliser practices during the "dry" growing season on post harvest soil N conditions discussed above. In chapter 4 the consequences of this approach are calculated for the experimental field. Spatial variation is characterised using three representative profiles for three subareas within the field. The three representative profiles for the soil units are identified and tested for internal homogeneity, by considering temporal variability of nitrate leaching during the "wet" season. Carry-over effects of N fertiliser management during the growing season to the "wet" season were quantified for the three representative profiles by simulations for an 11 year period. Thus, a probabilistic interpretation of five fertiliser scenarios resulted in recommended site-specific N fertiliser applications that allow high levels of potato production without exceeding the legal threshold for nitrate leaching. The latter is expressed in terms of probability of exceedance. Possibilities for farm management to focus fertiliser application during the growing season on post harvest N are discussed. Reducing nitrate losses to the environment by means of lower N fertiliser rates, taking into account spatial and temporal variation, is feasible under the studied conditions for the experimental field without significant losses in potato production.