作者: S Smart , E Rowe , C Evans , D Roy , I Moy
DOI:
关键词:
摘要: • Updated and finalized empirical niche models (GBMOVE) have been produced for 318 bryophytes and 851 higher plants species. Initial testing on an independent dataset of paired soil and plants species presence data showed good performance for an arbitrary subset of 9% of species models. Performance was very good when models were solved using observed mean Ellenberg values but less good when solved using observed soil data. This was due to the uncertainty introduced by the calibration functions that translate measurements of soil pH,% C,% N and soil moisture into mean Ellenberg N, R and F indices.• A new set of calibration equations were produced using a multiple regression approach. These achieved a significant improvement in r-sqrd value when explaining mean Ellenberg indices. While these improved functions still reduce the accuracy of niche models, when used to translate soil conditions into species composition they are a vital link between dynamic soil model outputs and GBMOVE niche models.