Because the strength of correlation depends strongly on the timeperiod
of analyses and the degree of yield variability over that period, it is
difficult to directly compare the yield correlations reported here with
those found in prior studies (e.g., Esquerdo et al., 2011; Liu & Kogan,
2002). For example, Liu and Kogan (2002) correlated AVHRR Vegetation
Condition Index and Temperature Condition Index data with CONAB
state-level soybean yield anomalies for 1986 to 1995 and obtained different
levels of peak correlation. However, the timing of peak sensitivities
was similar to those in Fig. 4, with MT, GO and MS peaking during
the month of January, when the flowering and grain-filling stages
occur, and PR peaking a few weeks earlier and RS a few weeks later.