Variations in forest water fluxes at local scale: effects of stand properties and weather

This thesis discusses the importance of various factors in regulating transpiration in a boreal forest. It synthesises field transpiration measurements by sapflow methods in mixed and pure stands of Norway spruce (Picea abies (L.) Karst.) and Scots pine (Pinus sylvestris L.), and modelling with a physically based, soil-oriented model. Two techniques for measuring sapflow (the Granier method and the Cermak method) were compared. Natural temperature gradients, distance between the sensors and radial flow patterns were shown to be important in the Granier system. The Granier system measured consistently lower flow rates than the Cermak system. It was possible to adjust the relationship between sapflow density and sapflow index for the Granier system, leading to fluxes comparable to those obtained by the Cermak measurements. Sapflow density was studied in relation to stand characteristics; it was possible to approximate transpiration in a number of adjacent forest stands at a local forest scale, using multiple regressions based on sapflow density and stand properties. Sapflow density decreased with increasing age, height, diameter, LAI and with decreasing stem density. Simulations of transpiration during the dry year 1994 identified a threshold value for reduction of water uptake. Because of wide data dispersion, it was not possible confidently to determine a threshold value at which soil moisture began to reduce transpiration when the quotient between measured/potential transpiration was compared with effective saturation. Correlation analysis showed that spruce was more sensitive than pine to changes in soil water availability. The modelling study showed that the response of canopy conductance to the effect of different light and vapour pressure deficit had little influence on estimates of transpiration at a daily time resolution. Use of mean transpiration of several stands permitted robust parameterisation of the model. The problems involved when using sapflow for estimation and modelling of canopy conductance are discussed; it is questioned whether time-lagged sapflow is suitable as a surrogate for transpiration in such calculations. Finally, as an example of the application of ‘water knowledge’, measured transpiration was used to calculate net canopy assimilation of six stands, using the concept water-use efficiency.