Underestimation of potassium in forest dry deposition? – A simulation experiment in rural Germany

Abstract

Measurements of throughfall (TF) and wet deposition (WD) are a common method to assess nitrogen (N) and base cation (BC) deposition to forests. Using TF and WD, dry deposition (DD) is usually calculated with a canopy budget model (CBM) assuming similar BC to Na+ ratios in WD and DD. This assumption is especially uncertain for K+, since K+ is often bound to smaller particles compared to Na+. Here we asses this assumption by comparing the DD of K+ estimated with the CBM (DDKCBM) to the DD of K+ simulated with a process oriented DD model (“inferential model”, DDKINF). Simulation experiments were performed at two indicator forest stands (“virtual” broadleaved (BL) and coniferous (CF) forest) based on six years of daily PM2.5 and PM10 concentrations and weekly WD observations measured at the rural background research station Melpitz, Germany. On average, the K+:Na+ ratio in WD was 0.24 while the K+:Na+ ratio in DDINF was 0.43 (CF) and 0.40 (BL), respectively. Accordingly, DDKCBM would need to be multiplied by a correction factor of 1.77 (CF) and 1.66 (BL) to match DDKINF, with substantial variation between years (lowest correction factor: 0.98, highest correction factor: 3.89). However, applying the correction factors in CBM calculations at nearby ICP Forests monitoring sites had only little effect on total (WD+DD) deposition rates of N and BC. The results were robust against changes in the meteorological data used for the inferential model. Uncertainty arises from periods affected by presence of particles larger than 10 µm diameter, not covered by local measurements. A corresponding lower boundary estimate for the average underestimation of DDKCBM is a correction factor of 1.37 (CF) and 1.29 (BL). More work is required to assess to what extend the observed underestimation of DDKCBM is confirmed by other methods and at sites with different atmospheric conditions.