Multitemporal Landsat-MODIS fusion for cropland drought monitoring in El Salvador

Nguyen Thanh Son, Chi Farn Chen, Cheng Ru Chen, Mario Giovanni Molina Masferrer, Luis Eduardo Menjívar Recinos

Research output: Contribution to journalArticlepeer-review

9 Scopus citations

Abstract

This study aims to develop an approach to characterize cropland drought conditions in El Salvador, Central America. The data were processed for 2016–2017 through three main steps: (1) reconstructing MODIS land-surface temperature (LST), (2) Landsat-MODIS data fusion and (3) drought delineation using the temperature vegetation dryness index (TVDI). The results of LST reconstruction using the random forests (RF) indicated the median RMSE value of 0.5 °C. The fusion results achieved from the STARFM compared with the reference Landsat data revealed close agreement with the correlation coefficient (r) values higher than 0.84. The TVDI results verified with that from the reference Landsat data indicated r values of 0.85 and 0.75 for 2016 and 2017, respectively. The larger very dry area was observed for the 2016 primera season due to prolonged droughts. Approximately 11.5% and 10.7% of croplands were, respectively, associated with very dry moisture condition in the 2016 and 2017 primera seasons.

Original languageEnglish
Pages (from-to)1363-1383
Number of pages21
JournalGeocarto International
Volume34
Issue number12
DOIs
StatePublished - 15 Oct 2019

Keywords

  • drought
  • El Salvador
  • Landsat
  • MODIS
  • STARFM

Fingerprint

Dive into the research topics of 'Multitemporal Landsat-MODIS fusion for cropland drought monitoring in El Salvador'. Together they form a unique fingerprint.

Cite this