Dados do Trabalho

Título

SOYBEAN YIELD RESPONSE TO UNDETAILED SOIL LIMITS AND PRECIPITATION

Resumo

Soil maps for large regions are available only at small scales and hold aggregated spatial information about soil attributes. Since soil attributes interact with weather variability to affect crop yield annual variability, there is a question about how useful limits from undetailed soil surveys are to account for soil-weather interactions that impacts crop yield over large regions. Do soil limits surveyed at small scales retain enough spatial information to explain crop yield variability as weather covariates are controlled over large regions?  It was checked by using remote sensing crop modeling into a covariance analysis (ANCOVA) to evaluate the effect of soil limits on soybean yield while the effect of the precipitation was controlled. Annual soybean yield maps were obtained by a calibrated model parameterized with crop vigor from MODIS data from 2007 to 2016 over 54 municipalities in Rio Grande do Sul. The annual yield variability was then spatially grouped by soil limits as made available by IBGE at 1: 250 000 scale. In addition, GPM precipitation data for the summer crop season was processed to daily mean precipitation and taken as covariate in the ANCOVA model. Precipitation was 25% lower in dry years, 4.63mm/day and 6.17mm/day for dry and humid years respectively. LVdf showed the highest yield difference (3.5 bags.ha-1) between humid and dry years.  The lowest difference (0.5 bags.ha-1) between humid and dry years was seen on NBa. The effect size was 0.06 (medium) and 0.03 (low) respectively for precipitation and soil limits. It was a general experiment that may be improved by including crop phenology synchronization with the weather data and crop management information. The low effect size of the undetailed soil limits indicates limited practical applications of undetailed limits to regional crop yield estimation. More investments are necessary to increase soil map detail to support crop modelling in regional context.

Palavras-chave

crop yield, effect size and soil survey scale

Instituição financiadora

CNPq

Agradecimentos

CNPq