Assessing Drought Vegetation Dynamics in Semiarid Grass- and Shrubland Using MESMA

Drought intensity and duration are expected to increase over the coming century in semiarid western United States due anthropogenic climate change. Historic data indicate that megadroughts this region have resulted widespread ecosystem transitions. Landscape-scale monitoring with remote sensing can help land managers track these changes. However, special considerations required: traditional vegetation indices such as NDVI often underestimate cover systems short multimodal green pulses, extremely variable rainfall, high soil fractions. Multi-endmember spectral mixture analysis (MESMA) may be more suitable, it accounts for both non-photosynthetic To determine suitability of MESMA assessing drought dynamics US, we test multiple endmember selection model parameters optimizing classification fractional (GV), (NPV), (S) grass- shrubland central New Mexico. Field spectra dominant species were collected at Sevilleta National Wildlife Refuge six field sessions from May–September 2019. Landsat Thematic Mapper imagery 2009 (two years pre-drought), Operational Land Imager 2014 (final year drought), 2019 (five post-drought) was unmixed. The best fit had levels agreement reference plots all three classes, R2 values 0.85 0.67 0.74 respectively. Reductions NPV increases GV S observed on landscape after event, persisted five a return normal rainfall. Results successfully applied changes relative fractions grass