COS 311-1 - Can remote sensing explain variation in functionally-similar pest densities across space?

Abstract:

Background:

The value of land-use intensity as a predictor of insect pest abundance has been hotly debated, but functional variation between insect species and their host environment has made broad comparisons difficult. Comparisons across space of functionally similar species in the same crop are needed. Miridae are major pests of cotton crops worldwide. They cause yield loss by piercing cotton squares and damaging boll set. Mirids are known to move large distances. Natural history observations suggest that homogenous agricultural landscapes result in earlier migration into cotton fields and increased overall abundance. Historic monitoring data of mirid pest density in cotton fields are used in this analysis from three different continents between 2000 and 2015 consisting of 4536 field-years in California, 1367 field-years in Spain and 373 field-years in Australia. We quantify the effect of direct measures of landscape productivity, derived from satellite remote sensing data, on mirid bug abundance in cotton. We test whether vegetation productivity and phenology (e.g. timing of dormancy) on the landscape scale are more important than precipitation and non-crop area for explaining variation of pest density. We test whether these metrics are equally important for predicting mirid bug density in cotton across three continents.



Results:

Our cross-continent decadal data support previous location-specific findings that higher early season mirid abundance results in lower cotton yield. Early season mirid abundance was compared to EVI index and vegetation phenology metrics at the 10km, 20km and 30km scales. Early-season mirid abundance in cotton fields declines with later landscape dormancy and higher EVI area in the non-crop habitat at the 30km scale. The proportion of non-crop habitat does not explain early-season mirid density in cotton. Phenological vegetation metrics are significantly correlated and show similar patterns, though the timing of dormancy is the most predictive for early-season mirid abundance. These results have important implications for landscape management decisions and in the context of global climate change, which often benefits agricultural pests over their predators. Furthermore, it shows the ecological value of earth observations in predicting invertebrate abundance and movement even in complex landscapes.