Field validation of soil solute profiles in irrigated cotton

Management of water and fertilizer N are important aspects of cotton production in the desert Southwest. GOSSYM, a cotton growth simulation model, has been used extensively to manage these inputs. Our objectives were to further validate GOSSYM by comparing model-simulated and measured soil NO₃/^--N profiles, to evaluate GOSSYM's potential as a management tool under irrigated growing conditions in the desert part of the U.S. Cotton Belt, and to address questions about the way GOSSYM simulates NO₃/^--N movement through the soil profile in relation to irrigation water management (which in turn affects prediction of plant growth and development). We compared measured profiles of NO₃^--N with GOSSYM-simulated profiles. Soil profile samples were obtained from an existing N-management field study, a split-plot within a randomized complete block design. Mainplots were upland and pima cotton (G. hirsutum L. cv. DPL 5415 and G. barbadense L. cv. Pima S-7, respectively). Subplots were a check (0 fertilizer N) and three other N-management strategies. The cotton was grown on a Casa Grande sandy loam [fine-loamy, mixed, hyperthermic Typic Natrargid (reclaimed)] near Maricopa, AZ, in 1994 and 1995. Fertilizer N rates ranged from 0 to 350 kg ha^-1 in 1994 and 0 to 392 kg ha^-1 in 1995. Soil samples taken to a depth of 120 cm in 30-cm increments were analyzed for NO₃/^--N. Comparisons of simulated and actual NO₃/^--N profiles revealed tendencies in GOSSYM to overestimate NO₃/^--N leaching out of the effective rooting zone, resulting in simulated N stresses midseason. When GOSSYM simulated an N stress, between 50 and 75% of the simulated soil NO₃/^--N values were greater than the measured values, yet the simulated N stress still occurred. This indicates possible limitations in GOSSYM's ability to adequately predict N uptake by plants. The dynamic soil N portion of the model needs further refinement, particularly for cotton production under irrigated desert conditions.