Is it possible to improve the general applicability and significance of empirical relationships between abiotic conditions and vegetation by harmonization of temporal data?
Three datasets of vegetation, recorded after periods with different meteorological conditions, were used to analyze relationships between soil moisture regime (expressed by the mean spring groundwater level –MSLt calculated for different periods) and vegetation (expressed by the mean indicator value for moisture regime Fm). For each relevé, measured groundwater levels were interpolated and extrapolated to daily values for the period 1970–2000 by means of an impulse-response model. Sigmoid regression lines between MSLt and Fm were determined for each of the three datasets and for the combined dataset.
A measurement period of three years resulted in significantly different relationships between Fm and MSLt for the three datasets (F-test, p < 0.05). The three regression lines only coincided for the mean spring groundwater level computed over the period 1970–2000 (MSLclimate) and thus provided a general applicable relationship. Precipitation surplus prior to vegetation recordings strongly affected the relationships.
Harmonization of time series data (1) eliminates biased measurements, (2) results in generally applicable relationships between abiotic and vegetation characteristics and (3) increases the goodness of fit of these relationships. The presented harmonization procedure can be used to optimize many relationships between soil and vegetation characteristics.