Back to the fields? Increased agricultural land greenness after a COVID-19 lockdown

Abstract

In response to the 2020 COVID-19 pandemic, policymakers worldwide adopted unprecedented measures to limit disease spread, with major repercussions on economic activities and the environment. Here we provide empirical evidence of the impact of a lockdown policy on satellite-measured agricultural land greenness in Badung, a highly populated regency of Bali, Indonesia. Using machine learning and satellite data, we estimate what the Enhanced Vegetation Index (EVI) of cropland would have been without a lockdown. Based on on this counterfactual, we estimate a significant increase in the EVI over agricultural land after the beginning of the lockdown period. The finding is robust to a placebo test. Based on evidence from official reports and international press outlets, we suggest that the observed increase in EVI might be caused by labour reallocation to agriculture from the tourism sector, hardly hit by the lockdown measures. Our results show that machine learning and satellite data can be effectively combined to estimate the effects of exogenous events on land productivity.

Delving into Geospatial Data Services: Monitoring Earth for Covid-19 Impact Measure and Decision Making

“…Their findings help evaluate the impact of Covid-19 on China's agricultural sector, which is promotive to formulating effective agricultural policy and ensuring food security. Hammad et al [25] provided empirical evidence of the after-effects of the lockdown policy on agricultural land greenness in Bali, Indonesia, using ML and SRS data and resulting in an Enhanced Vegetation Index (EVI) croplands. However, [26] showed that India had a good cropping season during the lockdown and post-lockdown periods.…”

Application of Probabilistic Machine Learning to the prediction of Remotely Sensed Vegetation Health

“…Scientists have used Machine Learning and Satellite Imagery to monitor tropical forest carbon stocks (Csillik et al, 2019), to identify and classify the direct drivers of deforestation (Irvin et al, 2020) and to estimate the effect of public policies on agricultural productivity (Hammad et al, 2021). Machine Learning algorithms provide unprecedented and promising results thanks to their ability to find patterns underlying the complex nonlinear relations that characterised environmental variables.…”