Application of spectral analysis to meteorological and soil solution chemistry data

One of the longest-running forest ecosystem studies in Europe was establish ed at the Selling region in central Germany. It provides a unique database covering more than 25 years of meteorological, throughfall, and soil soluti on chemistry data. It determines input to the ecosystem at canopy level and infiltration into the soil. These are important parts of water and element flux. Application of time series analysis can detect structures and dynami cs in the long-term time series, which are not apparent in the original dat a. Our study investigates these dynamics, especially their seasonal pattern s using spectral analysis. While some meteorological input parameters such as temperature and rain chemistry show pronounced annual seasonality, throu ghfall data of spruce and beech indicate a transformation of the input sign al. Chemical parameters of spruce throughfall behave either half-annual sea sonal (NO3-N, SO4-S, Ca), or, as chloride, the time series contains both an nual and half-annual seasonality. No element of beech throughfall has disti nct seasonal behaviour. During infiltration into sail, the next step of flu x, most seasonality disappears. Influences of seasonal inputs can he observ ed down to 0cm (under humus layer) or at mast 10cm under beech forest. Only potassium under spruce forest still shows strong seasonality at 100cm soil depth. This appears to he coupled with a biological process. (C) 1999 Else vier Science Ltd. All rights reserved.