A novel disease sensitive index (DSI) for monitoring early maize leaf disease using PROSPECT-D and LESS models

Maize leaf diseases have a significant impact on global industry and food security. Monitoring these diseases on a large scale with non-destructive remote sensing technology is a promising approach to mitigating their impacts. However, monitoring early maize leaf diseases at canopy scales is still challenging. Effective disease-specific indices have been developed at leaf scales, but the differences in planting density and cultivars may impact their performance for early crop disease monitoring at canopy scales. Here, we hypothesize that an index sensitive to leaf pigments and insensitive to plant structural parameters benefits early disease monitoring based on previous studies and field measurements. We analyzed the effects of leaf pigments and plant spacing (PS) on canopy spectral reflectance using PROSPECT-D and LESS models and found that the red-edge band (717 ± 5 nm) and green band (560 ± 10 nm) were more sensitive to leaf pigments than to PS. To better monitor early maize leaf diseases, we presented a new index, namely the disease sensitive index (DSI), which showed a high sensitivity to changes in leaf pigments, especially chlorophyll content, but a low sensitivity to PS. Our results demonstrated that the DSI was extremely statistically significant in differentiating rust severity, especially between low disease severity (p < 0.001). Furthermore, the DSI showed improved performance when quantifying the disease index (DI) across different years and sites (R2 = 0.69 in the Xinxiang dataset in 2021; R2 = 0.62 in the Yuanyang dataset in 2022). The time-series response patterns of vegetation indices (VIs) indicated that the DSI can detect disease onset as early as 10 days after infection (DAI 10). Our findings highlight the potential of multispectral remote sensing applications such as the DSI for monitoring early maize disease in precision agriculture.