Spatially Explicit Crop Model to Support the Prediction of Attainable Yield and Management Decisions in Sugarcane Fields

Abstract:

Brazil is the main sugarcane producer in the world, however due to the increasing demand for bioenergy and sugar, there is a need for improvements in management strategies. Crop models associated with Geographic Information System (GIS) could describe the spatial yield distribution, and consequently support management decisions. Thus, our goal was to provide sugarcane production in a 30 m grid resolution using DSSAT-CANEGRO (Decision Support System for Agrotechnology Transfer) to assist in agriculture practices. The study site was conducted at a 183-ha farm, heterogeneous in terms of soil and relief, located at southeastern Brazil. The steps of the methodology are as follows: a) 182 soil samples were collected in a regular grid, at surface (0-20 cm) and subsurface (80-100 cm) soil layers. Later, samples were analyzed for particle size and chemical attributes; b) We applied mass preserving equations to harmonized the soil data in regular depth intervals, with equal area-quadratic splines; c) Pedotransfer functions were performed to obtain hydrological soil properties; d) We used a geostatistical approach to spatialize soil attributes (Ordinary kriging); e) Daily climate data were obtained from a nearby weather station; f) We simulated a sugarcane planted in October with harvest completing 12 months; g) DSSAT was executed in batch file mode, where each file represented a cell grid. Yield prediction represented by fresh cane (SMDH), ranged from 73.9 to 139.3 t ha^-1 (Figure 1a). Higher values of SMDH were observed in regions of Nitisols, since presented more water availability to crop growth, being a key factor to sugarcane development. In addition, shallow soils (Leptosols and Cambisols) demonstrated susceptibility to water stress along the development cycle (Figure 1b). These preliminary results allow an understanding of yield spatial variability, which would enable to plan the agriculture calendar, irrigation and systematic delimitation of harvest areas at farm level.