Investigation of influential environmental and climatic determinants on COVID-19 spread in India to formulate a sustainable pandemic response.

The COVID-19 pandemic has highlighted the need for a Sustainable Pandemic Response Strategy (SPRS), driven by scientific research and engineering principles. This study focuses on Environmental and Climatic Determinants (ECDs) that may influence the occurrence pattern of infectious diseases. The objective of SPRS is to develop a climate-resilient framework for infectious diseases using Earth Observation (EO) data. ECDs were derived from EO data during the COVID-19 study period in India, spanning 1094 days (January 3, 2020, to December 31, 2022). A Convergent Search - Add or Eliminate (CS-AE) algorithm was developed for the investigation of complex association between ECDs and disease occurrence patterns. This algorithm identifies the most influential ECDs in the spread of COVID-19 in India, categorizing them as Determinants of Concern (DOC) or Determinants of Interest (DOI). Shortwave Downward Radiation (SDR) was identified as a DOC, showing a strong correlation ( = 0.9525) with COVID-19 spread. Granger causality analysis was conducted to support the classification of SDR as a Determinant of Concern (DOC). The results confirmed a temporal causal relationship between SDR and disease spread. During the first pandemic wave, significant causality was observed at lags of 2 to 7 days, with the strongest effect at lag 6 ( = 0.001), while in subsequent waves, significance was found across lags of 1 to 6 days. The seasonal effect of SDR and the three pandemic waves in India were observed through a radar chart, illustrating the temporal causal relationship between SDR and COVID-19 spread. The algorithm shows the note of a significant role by SDR in surface and air temperature ( = 0.9525;  = 0.9942) and influences other ECDs which are categorized as DOI. Hence, the proposed CS-AE algorithm provides a robust tool for identifying the most influential ECDs in the spread of infectious diseases, provided the datasets are time-series based.