Passive and Active Remote Sensing Data as Indicators of Vegetation Condition in Dry Woodland

Abstract

An important challenge is getting to know the condition of vegetation in dry woodlands, fragile ecosystems with high ecological value, to propose conservation strategies. Our objectives were: (1) to quantify on-the-ground vegetation composition and structure in dry Ramorinoa girolae-dominated woodlands on three study sites, using (A) field-based methods and (B) passive and active remotely sensed data at multi-scale and (2) to assess whether the integration of different remote sensors allows for a better estimation. We recorded field-based data on 36 plots of 50 × 30 m distributed in three sites (Ischigualasto n = 16; Valle Fértil n = 10; Private land n = 10). On each plot, we estimated the standard deviation (SD) of passive (EVI—Enhanced Vegetation Index—and SATVI—Soil Adjusted Total Vegetation Index) and active imagery data (SD of VV and VH polarisation) at multi-scale. The results of this approach show that passive and active remote sensing data were good indicators of almost all field-based data recorded at stand scale, except for richness, which was not related to remote sensing data. Shrub biomass, tree biomass and abundance of R. girolae were better explained by multi than single-sensor models. Only the variance in canopy cover of R. girolae was explained by single-sensor models. This work expands our understanding of the relationship of field-based data with both passive and active remotely sensed data, providing evidence that they could be used as a proxy for vegetation structure in dry R. girolae-dominated woodlands.