Evapotrantaspiration is a crucial part of the hydrological cycle but few ground observatories for the Tibetan Plateau exist. In this study, we present lysimeter measurements from the growing season during seven years at a remote field location on the Tibetan Plateau. The measurements show rates between 2.5 and 3 mm center dot d(-1) during the warmer months from June to August, dropping to 2 to 2.5 mm center dot d(-1) in September. This results in a total volume of evapotranspiration of approximately 300 mm center dot yr(-1) for the months from June to September. The inter-daily variability is however large, and comparison to meteorological variables suggest that this is largely driven by radiation and humidity. Data for a single season from a nearby flux tower allows us to compare the two common measurement methods for evapotranspiration in the field, showing an overall good agreement between the approaches. We also tested commonly applied models used to estimate evapotranspiration rates, namely the FAO-Penman-Monteith (PM) and the Priestly-Taylor (PT) model, which both make use of radiation data as well as the simpler Hargreaves-Samani (HS) and Rohwer (R) models which only need air temperature and wind speed as input. The most data intensive model (PM) has the lowest root mean square error (RMSE) (1.36 mm center dot d(-1)) and the mean bias error (MBE) (-0.05 mm center dot d(-1)) and reproduces the daily variability generally well. The much simpler HS model performs slightly worse (1.38 and 0.35 mm-d(-1)), but fails to reproduce the variability, due to its lack of information of local radiation and humidity data. Our results are in line with large scale estimates of evapotranspiration for the cold and arid region, provide a first long time series of in-situ measurements from a high elevation site and suggest that both the PM and HS models are appropriate when no direct measurements are available.

Russian boreal forests represent the largest forested region on Earth and comprise one-fifth of the world's forest cover. The two most common genera in Siberia are Larix and Pinus, which together cover more than 80% of the region's forested area. One observable ongoing effect of climate warming is that natural populations of Siberian larch are gradually being replaced by Scots pine. The present work focuses on comparing effects of environmental variables on sap flow density in two even-aged stands of Larix sibirica and Pinus sylvestris. While the two study stands were identical in age (49 years) with similar basal areas and leaf area index, they exhibited very different transpiration rates and response mechanisms to environmental signals. Stand water use was higher for larch than it was for pine, even though transpiration for deciduous larch trees occurred over shorter time periods. The cumulative annual transpiration of the larch stand was 284 +/- 4 mm measured over two consecutive growing seasons (2015-2016), while for pine this was 20% lower. Seasonal transpiration accounted for 50% and 40% of the reference evapotranspiration and 91% and 67% of growing season precipitation for larch and pine, respectively. Water stored in soil provided an important source of water for transpiration, observed as roughly 100 mm, which was then replenished from snowmelt the following spring. The greatest difference between two species related to how well they controlled transpiration, notably in the context of high vapor pressure deficit; under these conditions, pine maintained greater control over transpiration than larch. For all soil moisture levels measured, larch transpired more water than pine. Importantly, our results point to potential future effects of global warming, most notably an increasing decline of larch forests, changes in the ratio between latent and sensitive heat fluxes, and significant modifications in ecosystem water availability.