Nitrogen leaching has caused a growing societal concern over N fertilizer impact on water quality. One way to decrease nitrogen loss through leaching is to adjust fertilizer inputs to site-specific conditions. This study was conducted to investigate spatial variability of NO3 leaching parameters on a 5 ha commercial wheat field (Typic Ustifluent) located 25 km north of Tokat, Turkey, for the purpose of dividing the field into small cells in which application rates can be kept constant. NO3 leaching parameters were calculated using the monthly analysis version of computer program NLEAP (nitrate teach and economic analysis package) on a regular grid spacing of 25 m, and semi-variogram for each parameter was calculated using the computer program GEAOES. The values for parameter NL (nitrate leached) were between 24.64 (low) and 77.28 kg ha(-1) (medium), for NAL (nitrate available for leaching) 42.46 (low) and 274.40 kg ha(-1) thigh), and for MRI (movement risk index) 0.28 (low) and 0.35 (medium). Values for parameter ALRP (annual leaching risk potential) varied from high tinder = 4) to moderate tinder = 3). A moderately significant correlation (r = 0.54, P < 0.01) was found between measured and model-estimated values for the parameter NAL, indicating that the NLEAP model adequately simulated the NO3 leaching in the study area. Values for range were 360 m for NAL, 350 m for NL and 180 m for MRI, and nugget effect was 0.72 for MRI, 0.45 for NAL and 0.25 for NL, and mean correlation distances (MCD) were 145 m for NAL and 61 m for NL. Although, the spatial patterns for the parameters NAL and NL were similar, the upper cell limit for parameter NAL was higher than two times that of parameter NL, suggesting that calculation of input for continuous control of nitrogen application rate in a variable rate nitrogen fertilizer application program be based on the spatial pattern of NL but not on that of NAL. (C) 2001 Elsevier Science B.V. All rights reserved.