Detection of field significant long-term monotonic trends in spring yields

Trend analysis is a frequently used tool in hydrology and climatology for the identification of long-term changes. However, studies are usually only oriented on local trends. This paper rather focuses on the spatial application of trend analysis in groundwater data. For this purpose, a modification of the Mann-Kendall test was developed, based on the trend-free pre-whitening approach. This method was successfully tested on 157 series of yields from headwater springs collected in Czechia during the 1971–2007 period. The analysis was done separately for year, each season and each month. Field significant trends in spring yields were identified in hydrogeological regions. The results showed that the field significant trends are outnumbered when cross-correlation is not taken into account. In the case of annual series, 4 of 18 hydrogeological regions investigated showed a significant decreasing trend after corrections for cross-correlation, compared to 12 regions with field significant trend when not considering cross-correlation.