Halabuk A., Gerhatova K., Kohut F., Ponecova Z., Mojses M.: Identification of season-dependent relationships between spectral vegetation indices and aboveground phytomass in alpine grassland by using field spectroscopy. Ekologia (Bratislava), Vol. 32, No. 2, p. 186-196, 2013.
Spectral characteristics of alpine grasslands across the vegetation season (from May to September) are presented. The results are based on three year field spectroscopy monitoring of acid, nutrient poor grasslands at Kraľova hoľa research site, Low Tatras, Slovakia. Relationships between commonly used spectral vegetation indices (VIs) and field-based estimation of aboveground green phytomass (AG B) were analysed. Finally, season-dependent regression models were created in order to allow spatially extensive non-destructive monitoring of AG B. Spatial-temporal dynamics of background and standing litter markedly affect seasonal variations of relationships between VIs and AG B and predictability of the regression models. Because of a high proportion of litter during the whole season, this was a plant water-sensitive normalized difference water index (NDWI), which dominates as the predictive variable in the regression models across the whole season; except June, where chlorophyll absorption sensitive in normalized difference vegetation index (NDVI) performed the best (R2 = 0.57; rel. RMSE = 34%). However, the accuracy of the models was quite low (May: R2 = 0.45; rel. RMSE = 49%; July: R2 = 0.47; rel. RMSE = 26%; August: R2 = 0.13; rel. RMSE = 31%; September: R2 = 0.53; rel. RMSE = 40%).