Long-term (1990 – 2015) water balances of Norway spruce and Scots pine stands in Finland

The balance between incoming precipitation (rainfall and snowfall) and outgoing evapotranspiration (ET), runoff and drainage to and from an ecosystem plus changes in soil moisture storage and the water equivalent of the snowpack is known as the water balance. A dominating feature of the water balance in the boreal zone is snowpack accumulation over winter and the spring snowmelt, both of which are affected by forest. In Finland, there are strong north-south gradients in the amount of precipitation, the proportion of rainfall and snowfall and temperature, and therefore latitudinal differences in the water balances components can be expected. Furthermore, the large canopy and deeper rooting of trees, together with the presence of a permanent ground vegetation cover, result in significant differences in interception, infiltration and water balance outputs of forests compared to other forms of land-use. Because of morphological and ecophysiological differences between the trees species, the water balance of Norway spruce and Scots pine dominated forests can be expected to differ. Determining the water balance of forest ecosystems across Finland would, therefore, help in assessing the hydrological ecosystem services provided by forests and form a basis for examining the effects of climate change and forest management on the water balance. This study aimed to compute the daily water balance of six Norway spruce, and three Scots pine dominated mature forest stands (plots) located throughout Finland over a 26-year study period (1990-2015). It was hypothesized that the various water balance components would systematically vary with latitude, a surrogate for climate, and differ between spruce and pine stands. The daily version of the water balance model “WATBAL” developed by Mike Starr (University of Helsinki, Dept. Forest sciences) was used for this study. The model requires daily meteorological data (precipitation, temperature, global radiation), stand parameters (canopy cover, rooting, crop coefficient), soil parameters (including infiltration coefficient, soil moisture contents at permanent wilting point, field capacity and saturation, and two soil moisture parameters for a plant available water content function). Six of the plots had soil developed in till and 3 plots had soil developed in sorted glaciofluvial deposits. Plot meteorological data for 1990-2015 was derived using spatially interpolated gridded data. If the daily air temperature was ≤0°C, any precipitation was assumed to be snowfall. The stand and soil parameters were derived from data collected from the 9 study plots by Luke (formerly Metla). The nine plots belong to the Finnish network of ICP-Forest level II plots that have been established throughout Europe. Pedotransfer functions (PTF) based on soil texture and organic matter contents were used to derive initial values for the soil hydraulic parameters. Time domain reflectometry (TDR) measured soil moisture data was available for 7 of the plots and, after carrying out careful quality control and rejection of outliers, used for calibration of modelled soil moisture and optimization of soil hydraulic parameters for those plots. Optimization was carried out using the non-linear Marquardt regression method. Goodness-of-fit for soil moisture was evaluated using correlation and R2 values from linear regression. After computing the daily water balance with the WATBAL model (using optimized soil hydraulic parameter values for the 7 plots and initial PTF values for the remaining 2 plots) the long-term mean annual and mean daily water balance components (with a 7-day moving average smoothing) were calculated. The water balances were computed for the humus layer plus 0-40 cm soil layer, which, based on literature, would have included most if not all of the roots. The dependence of the mean annual water balance components on latitude was evaluated using correlation analysis and linear regression, and the effect of tree species was tested for using the t-test on pairs of spruce-pine plots located close to each other. The raw TDR data was found to contain a considerable amount of gaps and erroneous (too high) values, often associated with the spring snowmelt. Optimization of the soil hydraulic parameters using the measured soil moisture contents calculated from the “cleaned” TDR data for the snow-free period resulted in a highly significant (p<0.001) Pearson correlation of +0.85 (R2 = 0.75) for the fit between measured and modelled soil moisture contents calculated across all 7 plots. The correlations for the individual plots were also highly significant. Based on the optimized WATBAL output, the fraction of plot mean annual precipitation as snowfall ranged from 20 to 29%. Corresponding ranges for ET, drainage and runoff were respectively 33 to 57%, 24 to 42%, and 18 to 25%. The mean annual water balance components were found to be significantly correlated to latitude, reflecting trends in precipitation and temperature. Evapotranspiration decreased with increasing latitude while maximum snow-on-ground, snowmelt and associated runoff increased with increasing latitude. Spruce mean annual ET was 9% higher than pine in one of the paired plot sets and 37% higher in the other set. For drainage, pine was 15% greater than spruce in one of the paired plot set and 74% higher in the other set of paired plots. There were no significant differences between spruce and pine plots for snowmelt and runoff. Variation around these trends were related to differences in soil hydraulic properties among the plots which, in turn, were related to differences in parent material and soil texture. The overall conclusion from this study was that the daily water balance of the forested plots could be realistically modelled using such a relatively simple water balance model as WATBAL. The importance of spatially representative and accurate soil moisture measurements for model calibration purposes was highlighted. While the importance of snowfall on the water balance increased northwards regardless of tree species, evapotranspiration was determined by both latitude and by species. Climate change can therefore be expected to have a significant impact on the water balance of Finnish forests resulting in environmentally important changes in leaching and runoff.