Modeling cotton growth and yield response to irrigation practices for thermally limited growing seasons in Kansas

Highlights Later planting and greater location elevation or latitude decreased seasonal growing degree days and cotton yield, Higher irrigation capacity (rate) usually increased lint yield that was probably due to increased early boll load, Strategies for splitting land allocations between high irrigation rates and dryland did not increase production, Cotton may reduce irrigation withdrawals from the Ogallala Aquifer, but the Kansas growing season limits production. Abstract. The western Great Plains precipitation averages about 450 mm, varying little with latitude and providing from 40-80% of the potential crop evapotranspiration (ETc). Supplemental irrigation is required to fully meet crop water-demand, but the Ogallala or High Plains Aquifer source is essentially non-recharging south of Nebraska. Pumping water from this aquifer draws down water tables, leading to reduced water availability and deficit irrigation to produce an alternate crop like cotton [Gossypium hirsutum (L.)] with a lower peak water demand than corn [Zea mays (L.)]. Our objective was to compare simulated cotton yield response to emergence date, irrigation capacity, and application period at three western Kansas locations (Colby, Tribune, Garden City) with varying seasonal energy or cumulative growing degree days (CGDD) and compare split center pivot deficit irrigation strategies with a fixed water supply (i.e., where portions of the center pivot land area are managed with different irrigation strategies). We used actual 1961-2000 location weather records with the GOSSYM simulation model to estimate yields of cotton planted into soil at 50 % plant available water for three emergence dates (DOY 145, 152, and 159) and all combinations of irrigation periods (0, 4, 6, 8, and 10 weeks beginning at first square) and capacity (2.5, 3.75,and 5.0 mm d-1). Simulated lint yield and its ratio to ETc or the water use efficiency (WUE) consistently decreased with delayed plantings (emergence) as location elevation or latitude increased due to effects on growing season CGDD. Depending on location, simulated cotton lint consistently increased (P=0.05) for scenarios with increasing irrigation capacity that promoted greater early season boll load but not for durations exceeding 4 to 6 weeks probably because later irrigation and fruiting did not complete maturation during the “short” growing season. Cotton WUE generally increased with greater yields resulting from earlier emergence and early high capacity irrigation. We calculated lower WUE where irrigation promoted vigorous growth with added fruiting forms that delayed maturation and reduced the fraction of open bolls. The irrigation strategy of focusing water at higher capacities on a portion of the center pivot in combination with the dryland balance did not increase net yields significantly at any location since the available seasonal energy limited potential crop growth and yield response to irrigation. The overall net lint yield for focused irrigation strategies at the southwest Kansas location (Garden City) was, however, numerically larger. Based on lint yields simulated under uniform or split center pivot deficit irrigation, we conclude that cotton is poorly suited as an alternative crop for central western and northwestern Kansas because of limited growing season CGDD.