Indicator of flower status derived from in situ hyperspectral measurement in an alpine meadow on the Tibetan Plateau

Flowering status including flowering date and flower amount could reflect ecological process in assessing plant phenological response to global warming. However, little information is available so far for monitoring flowering status through remote sensing. To provide an ecological indicator for monitoring plant phenology from remotely sensed data, we conducted a field survey in an alpine meadow on the Tibetan Plateau where flower color in July is dominantly yellow due to flowering of Halerpestes tricuspis (Ranunculaceae). We used flower coverage to indicate the flowering status of this species and proposed a flower index derived from in situ hyperspectral data (HFI) to estimate the flower coverage. Results demonstrate that the flower coverage of H. tricuspis can be estimated with high accuracy from the hyperspectral measurements. The indicating ability was further improved when the flower coverage was higher than 0.10 or the fractional coverage of soil was low or known in advance. A simulation also shows that a quadrat or pixel with flower coverage higher than 0.066 can be detected with existence of flower by HFI if soil fraction is less than 50%. These results indicate that HFI is applicable for estimating flower coverage of this species from hyperspectral measurement. The study suggests that the hyperspectral remote sensing technique can be applied for monitoring flowering status, and therefore the technique can provide an important ecological indicator for monitoring plant phenology. (C) 2008 Elsevier Ltd. All rights reserved.