Climate and soil-based constraints to rainfed cotton yield in the Northern Territory, Australia – A modelling approach using APSIM-OZCOT

Abstract

Cotton production in Australia is expanding to the northern semi-arid tropical regions as a means to droughtproof the industry. The majority of expansion is expected to come from rainfed farming systems. However, little is known on tropical rainfed cotton, particularly on climate and soil-based constraints that determine cotton production. Using the Northern Territory (NT) as case study, we investigated the effects of between and intra seasonal rainfall variability and key soil types on soil moisture dynamics,timely operations, lint yield, and industry viability. We examined several regions with greatest potential for rainfed agriculture in the NT, including sites in the Daly Basin, Mataranka and Larrimah. To this purpose we applied a locally calibrated and validated cotton growth model (APSIM-OZCOT) using weather files sourced from the SILO database of Australian climate data, and soil profile parameters for three red Kandosols: Blain, Oolloo and Tippera. Our study shows that simulated median lint yields ranged from 570 to 996 kg ha−1, with larger yields achieved on soils that hold more moisture (Blain > Oolloo > Tippera), and at locations that received greater amounts of seasonal rainfall. Planting window largely governed soil-moisture availability and thus sowing opportunities. An early planting (1–15 December) increased yields (greater available soil-moisture) but reduced planting opportunities due to insufficient mulch cover at planting. A late planting (15 Jan to 1 Feb) reduced cotton yields due to less seasonal rainfall, but increased the number of sowing opportunities in the drier regions since more time was allowed to establish sufficient cover crop. Availability of deep soil-moisture was important, particularly in the later crop phases when the roots reach deeper layers, and sufficient soil moisture is required to grow and finish bolls. Our assessment further highlights the potential impact of excessively wet conditions on cotton production. Short periods of intense rainfall (known as “burst” events); a typical characteristic of the monsoon, reduce cotton yield through waterlogging. Scaling our simulated yields to the potential area of suitable soils in the NT shows considerable benefits to the regional economy and should be considered when planning industry development, including the establishment of cotton gins. The insights gained from this study are relevant to industry developments in similar environments, such as in parts of Brazil, India, and Sahelian Africa