Going beyond the green: senesced vegetation material predicts basal area and biomass in remote sensing of tree cover conditions in an African tropical dry forest (miombo woodland) landscape
In sub-Saharan Africa (SSA), tropical dry forests and savannas cover over 2.5 million km2 and support livelihoods for millions in fast-growing nations. Intensifying land use pressures have driven rapid changes in tree cover structure (basal area, biomass) that remain poorly characterized at regional scales. Here, we posed the hypothesis that tree cover structure related strongly to senesced and non-photosynthetic (NPV) vegetation features in a SSA tropical dry forest landscape, offering improved means for satellite remote sensing of tree cover structure compared to vegetation greenness-based methods. Across regrowth miombo woodland sites in Tanzania, we analyzed relationships among field data on tree structure, land cover, and satellite indices of green and NPV features based on spectral mixture analyses and normalized difference vegetation index calculated from Landsat 8 data. From satellite-field data relationships, we mapped regional basal area and biomass using NPV and greenness-based metrics, and compared map performances at landscape scales. Total canopy cover related significantly to stem basal area (r 2?=?0.815, p?0.01) and biomass (r 2?=?0.635, p?0.01), and NPV dominated ground cover (>?60%) at all sites. From these two conditions emerged a key inverse relationship: skyward exposure of NPV ground cover was high at sites with low tree basal area and biomass, and decreased with increasing stem basal area and biomass. This pattern scaled to Landsat NPV metrics, which showed strong inverse correlations to basal area (Pearson r?=??0.85, p?0.01) and biomass (r?=??0.86, p?0.01). Biomass estimates from Landsat NPV-based maps matched field data, and significantly differentiated landscape gradients in woody biomass that greenness metrics failed to track. The results suggest senesced vegetation metrics at Landsat scales are a promising means for improved monitoring of tree structure across disturbance and ecological gradients in African and other tropical dry forests.