VALIDATING WATER USE AND YIELD ESTIMATES DERIVED FROM REMOTE SENSING AND CROP MODELLING FOR IRRIGATED SUGARCANE IN MPUMALANGA, SOUTH AFRICA

摘要

Accurate and timeous information on crop growth and crop water use are crucial to support the management of irrigated sugarcane production. These can be estimated from weather-based simulation models (e.g. the Canesim sugarcane model) and with remotesensing technology (such as the Surface Energy Balance Algorithm for Land - SEBAL), but the accuracy of these estimates needs to be determined before operational application. The aim of this study was to evaluate the accuracy of Canesim and SEBAL estimates of crop growth and water use by comparing it to ground measurements taken in thirteen sugarcane fields in Mpumalanga, South Africa. The Canesim model was used to simulate daily evapotranspiration (ET), canopy cover (CC), dry aerial biomass (ADM) and dry stalk mass (SDM) for each field using appropriate soil, weather, irrigation and crop input data. The SEBAL method was used to estimate, from satellite imagery and weather data, weekly totals of ET and dry biomass production. Weekly values of CC was also derived from satellite imagery. An algorithm was developed to estimate ADM, SDM and cane yield from remotely sensed data and temperature. This information was produced for 52 consecutive weeks from 3 November 2011 to 31 October 2012. CC was also measured in each field with a line quantum sensor at monthly intervals and ADM and SDM were determined from samples taken on three to four occasions. Daily ET was estimated in one field using the surface renewal (SR) technique. CC estimated from satellite imagery was much more accurate than that estimated by Canesim. SEBAL and Canesim ET estimates were similar with both exceeding estimates from SR measurements by about 8 mm/week. SEBAL estimates of stalk dry mass and cane yields were similar to that of the Canesim model, when the latter used measured soil water content data as input. An advantage of using a remote sensing technique is that it can provide spatial estimates at 30 m resolution of all variables reported in this study, while crop model estimatesare point based and cannot account for within-field variation. SEBAL data could be used to identify sugarcane areas with water deficit, slow growth or low yield at field, farm and regional levels, enabling corrective action to be taking early. Remotelysensed CC and SEBAL estimates of ET and biomass production could also enhance the accuracy of yield forecasts from models, at mill and field levels.