The integration of plant phenology and land use data to create a gis-assisted bioclimatic characterisation of bavaria, Germany

Background: Since phenology is a good bio-indicator of temperature the latter is often used in phenological analyses. However, whilst meteorological data are difficult to interpolate from point measurements at the desired resolution, region-wide digital data, e.g. elevation models or land cover data, are usually readily available. Aims: The potential of environmental variables, other than meteorological data, to create a bioclimatic classification of landscapes at the mesoscale was tested by the joint use of spatial data and the flowering dates of Forsythia suspensa at 70 phenological stations in southern Bavaria, Germany. Methods: A linear discriminant analysis was carried out to identify relevant land use variables that were correlated with phenology and, using these results, to regionalise the observed flowering dates within the framework of a Geographic Information System. Results: The generated map represented dates of onset at the regional scale, mostly influenced by the extent of impervious (hard, sealed) surfaces and forest, altitude and distance to urban areas. Thus, we overcame restrictions resulting from the difficulties of spatially interpolating available climatological data, and from the limited number of phenological datasets. Conclusions: We demonstrated that the selected variables were capable of adequately modelling regional bioclimatic zonation, and that phenology was a useful proxy of regional climate variation arising from both natural and anthropogenic factors.