Search

In summer 2018, Northern Europe experienced an extreme summer drought in combination with unusually high temperatures, which had a substantial impact on agricultural yields as well as on forest growth conditions in various ways. An earlier study, using ICOS RI (Integrated Carbon Observation Research Infrastructure) stations and other forest ecosystem stations in the Nordic region, shows that the drought dramatically decreased NEP in the southern Scandinavian and Baltic region, almost nullifying the carbon sinks in some of the forests [1]. Such severe conditions during a single year could be expected to influence a forest over following several years. Reduced tree storage of carbohydrates leads to a changed carbon allocation pattern in spring that may affect both the woody growth and pests' resistance. It is thus important to reveal the impact of such climatic events over a more extended period. This study aims at assessing the carry-over effects of the extreme weather conditions on the carbon and water fluxes and the forest growth to the years after the event. The analysis is based on measurement from the stations shown to be significantly affected by the drought through reduced carbon fluxes in 2018: the spruce forests Hyltemossa and Skogaryd and the mixed forests Norunda, Svartberget, Soontaga and Rumperöd. The ecosystem carbon and water fluxes will, together with tree-ring width data, be used to assess the carbon and water exchange and growth recovery in the years after the extreme 2018 drought (2019 and 2020) by comparisons to earlier normal years and extreme events.[1] Lindroth, A., et al. (2020): Effects of drought and meteorological forcing on carbon and water fluxes in Nordic forests during the dry summer of 2018 Phil. Trans. R. Soc. B37520190516 https://doi.org/10.1098/rstb.2019.0516

Tree killing by spruce bark beetles (Ips typographus) is one of the main disturbances to Norway spruce (Picea abies) forests in Europe and the risk of outbreaks is amplified by climate change with effects such as increased risk of storm felling, tree drought stress and an additional generation of spruce bark beetles per year[1]. The warm and dry summer of 2018 triggered large outbreaks in Sweden, the increased outbreaks are still ongoing and affected about 8 million m3 forest in 2020[2]. This is the so far highest record of trees killed by the spruce bark beetle in a single year in Sweden[2]. In 1990-2010, the spruce bark beetle killed on average 150 000 m3 forest per year in southern Sweden[3]. Bark beetles normally seek and attack Norway spruces with lowered defense, i.e. trees that are wind-felled or experience prolonged drought stress[4]. However, as the number of bark beetle outbreaks increase, the risk of attacks on healthy trees also increase[5]. This causes a higher threat to forest industry, and lowers the possibilities to mitigate climate change in terms of potential decreases in carbon uptake if the forests die[4,5]. Norway spruce trees normally defend themselves by drenching the beetles in resin[6]. The resin in turn contains different biogenic volatile organic compounds (BVOCs), which can vary if the spruce is attacked by bark beetles or not [4,6]. The most abundant group of terpenoids (isoprene, monoterpenes and sesquiterpenes), is most commonly emitted from conifers, such as Norway spruce[7,8]. The aim of this study was to enable a better understanding of the direct defense mechanisms of spruce trees by quantifying BVOC emissions and its composition from individual trees under attackTo analyze the bark beetles’ impact on Norway spruce trees a method was developed using tree trunk chambers and adsorbent tubes. This enables direct measurements of the production of BVOCs from individual trees. Three different sites in Sweden, with different environmental conditions were used for the study and samples were collected throughout the growing season of 2019. After sampling, the tubes were analyzed in a lab using automated thermal desorption coupled to a gas chromatograph and a mass spectrometer to identify BVOC species and their quantity.The preliminary results show a strong increase in BVOC emissions from a healthy tree that became infested during the data collection. The finalized results expect to enable better understanding of how spruce trees are affected by insect stress from bark beetles, and if bark beetle infestation will potentially result in increased carbon emission in the form of BVOCs.References[1] Jönsson et al. (2012). Agricultural and Forest Meteorology 166: 188–200[2] Skogsstyrelsen, (2020). https://via.tt.se/pressmeddelande/miljontals-granar-dodades-av-granbarkborren-2020?publisherId=415163&releaseId=3288473[3] Marini et al. (2017). Ecography, 40(12), 1426–1435.[4] Raffa (1991). Photochemical induction by herbivores. pp. 245-276[5] Seidl, et al. (2014). Nature Climate Change, 4(9), 806-810. [6] Ghimire, et al. (2016). Atmospheric Environment, 126, 145-152.[7] Niinemets, U. and Monson, R. (2013). ISBN 978-94-007-6606-8[8] Kesselmeier, J. and Staudt, M. (1999). Journal of Atmospheric Chemistry, 33(1), pp.23-88

Since 1750, land-use change and fossil fuel combustion has led to a 46% increase in the atmospheric carbon dioxide (CO2) concentrations, causing global warming with substantial societal consequences. The Paris Agreement aims to limit global temperature increases to well below 2°C above preindustrial levels. Increasing levels of CO2 and other greenhouse gases (GHGs), such as methane (CH4) and nitrous oxide (N2O), in the atmosphere are the primary cause of climate change. Approximately half of the carbon emissions to the atmosphere are sequestered by ocean and land sinks, leading to ocean acidification but also slowing the rate of global warming. However, there are significant uncertainties in the future global warming scenarios due to uncertainties in the size, nature, and stability of these sinks. Quantifying and monitoring the size and timing of natural sinks and the impact of climate change on ecosystems are important information to guide policy-makers’ decisions and strategies on reductions in emissions. Continuous, long-term observations are required to quantify GHG emissions, sinks, and their impacts on Earth systems. The Integrated Carbon Observation System (ICOS) was designed as the European in situ observation and information system to support science and society in their efforts to mitigate climate change. It provides standardized and open data currently from over 140 measurement stations across 12 European countries. The stations observe GHG concentrations in the atmosphere and carbon and GHG fluxes between the atmosphere, land surface, and the oceans. This article describes how ICOS fulfills its mission to harmonize these observations, ensure the related long-term financial commitments, provide easy access to well-documented and reproducible high-quality data and related protocols and tools for scientific studies, and deliver information and GHG-related products to stakeholders in society and policy.

In summer 2018, Northern Europe experienced an extreme summer drought in combination with unusually high temperatures, which had a substantial impact on agricultural yields as well as on forest growth conditions in various ways. An ongoing study, using ICOS and other forest ecosystem stations in the Nordic region, shows that the drought dramatically decreased NEP in the southern Scandinavian and Baltic region, almost nullifying the carbon sinks in some of the forests. However, some of the forests that not were exposed to the most extreme drought actually increased their NEP because of the high evaporative demand. Such severe conditions during a single year could be expected to influence a forest over several following years. Reduced tree storage of carbohydrates leads to a changed carbon allocation pattern in spring that may affect both the woody growth and the resistance to pests. It is thus important to reveal the impact of such climatic events over a longer period. This study aims at assessing the carry-over effects of the extreme weather conditions on the carbon fluxes and the forest growth to the year after the event, 2019. The base of the analysis will be eddy covariance data combined with tree ring time series from measurement stations that has been shown to be significantly affected by the drought through reduced carbon fluxes: the spruce forests Hyltemossa and Skogaryd and the mixed forests Norunda, Svartberget, Soontaga and Rumperöd. The eddy covariance and tree ring data will be used to assess the forest ecosystem carbon fluxes and growth recovery in 2019 by comparisons to earlier normal years and extreme events.

The Nordic region was subjected to severe drought in 2018 with a particularly long-lasting and large soil water deficit in Denmark, Southern Sweden and Estonia. Here, we analyse the impact of the drought on carbon and water fluxes in 11 forest ecosystems of different composition: Spruce, pine, mixed and deciduous. We assess the impact of drought on fluxes by estimating the difference (anomaly) between year 2018 and a reference year without drought. Unexpectedly, the evaporation was only slightly reduced during 2018 compared to the reference year at two sites while it increased or was nearly unchanged at all other sites. This occurred under a 40 to 60% reduction in mean surface conductance and the concurrent increase in evaporative demand due to the warm and dry weather. The anomaly in the net ecosystem productivity (NEP) was to 93% explained by a multilinear regression with the anomaly in heterotrophic respiration and the relative precipitation deficit as independent variables. Most of the variation (77%) was explained by the heterotrophic component. Six out of 11 forests reduced their annual NEP with more than 50 g C m−2 yr−1 during 2018 as compared to the reference year. The NEP anomaly ranged between −389 and + 74 g C m−2 yr−1 with a median value of −59 g C m−2 yr−1.

Europe is one of the world’s largest food producers, and climate change may pose a serious threat to food security in the region. In the present study, we assess the potential impact of climate change on the Colorado Potato Beetle (CPB), Leptinotarsa decemlineata (Say)—a severe pest of potato (Solanum tuberosum, L.). We also investigate the possible impact of climate change on the phenological development of potato. The main focus is on factors that may limit the northward expansion of the CPB, and the number of generations per year in areas where the insect pest is already present. These factors include lack of temperature sum for completed development before winter, and lack of food (i.e. potato) due to mismatches in insect-host plant phenological synchrony. We use a gridded observational dataset and an ensemble of bias corrected regional climate model data for the period of 1981–2099, representing RCP8.5, as input to a potato and CPB phenological model. The results show that in the future, CPB individuals with a low developmental threshold (+10 °C) can complete maturity of two generations per year before potato is harvested in most parts of Europe. A third generation of CPB may not be able to complete maturation due to lack of food in south and central Europe, while temperature becomes a limiting factor further north. In north-eastern Europe, the initiation of a first generation may be delayed due to lack of food in spring. CPBs with a high developmental threshold (+12 °C) will emerge later from winter hibernation, and food availability will therefore not be a problem in spring. However, individuals with a high developmental threshold face a greater risk of regeneration failure caused by harvesting of potato in autumn. The potential lack of food in autumn may also increase the strength of selection towards a low developmental threshold in northern populations. The combined analysis of CPB and potato phenology indicated that climate change can lead to increased pressure from the CPB in most potato growing areas.

Potato (Solanum tuberosum) is one of the main food crops in northern Europe, considered to be the fourth most important crop on a global scale after rice, wheat and maize. Climate change leading to longer growing seasons may call for adjustments in timing of planting and harvesting, and in this modeling study we assess potential effects of a warmer climate on potato crop phenology and temperature stress. A phenological potato model was parameterized with three planting dates to assess management impact on the timing of emergence and maturation of both early and late potato. Estimates on phenological development and occurrence of temperature stress were analyzed by comparing two developmental thresholds (0 °C and +2 °C) and three temperature response functions. The potato model was driven by observed gridded climate data and two sets of bias corrected climate model data, representing RCP4.5 and RCP8.5 for the period 1991–2100. The future simulations indicated that a warmer climate and earlier planting may move the timing of harvest up to 1 month earlier, however, potato emergence early in the year will be associated with an increased risk of frost damage in most parts of northern Europe. The areas of west Europe most prone to frost damage today may experience a risk of frost damage in response to climate change. The simulation of early potato development was sensitive to the setting of the developmental threshold, while late potato development was sensitive to the optimum temperature setting. While a linear temperature response function is essentially sufficient for current climate conditions in northern Europe, optimum and upper thresholds should be considered in climate change impact assessments. The potato model runs with temperature data corrected according to quantile-mapping indicated in general a slightly higher risk of temperature stress than the corresponding runs with temperature data corrected by linear scaling.

The terrestrial carbon and water cycles are intimately linked: the carbon cycle is driven by photosynthesis, while the water balance is dominated by transpiration, and both fluxes are controlled by plant stomatal conductance. The ratio between these fluxes, the plant water-use efficiency (WUE), is a useful indicator of vegetation function.WUE can be estimated using several techniques, including leaf gas exchange, stable isotope discrimination, and eddy covariance. Here we compare global compilations of data for each of these three techniques.We show that patterns of variation in WUE across plant functional types (PFTs) are not consistent among the three datasets. Key discrepancies include the following: leaf-scale data indicate differences between needleleaf and broadleaf forests, but ecosystem-scale data do not; leaf-scale data indicate differences between C3 and C4 species, whereas at ecosystem scale there is a difference between C3 and C4 crops but not grasslands; and isotope-based estimates of WUE are higher than estimates based on gas exchange for most PFTs.Our study quantifies the uncertainty associated with different methods of measuring WUE, indicates potential for bias when using WUE measures to parameterize or validate models, and indicates key research directions needed to reconcile alternative measures of WUE.

The effect of water availability on stand-level productivity, transpiration, water use efficiency (WUE) and radiation use efficiency (RUE) is evaluated for different willow clones at stand level. The measurements were made during the growing season 2000 in a 3-year-old plantation in Scania, southernmost Sweden. Six willow clones were included in the study: L78183, SW Rapp, SW Jorunn, SW Jorr, SW Tora and SW Loden. All clones were exposed to two water treatments: rain-fed, non-irrigated treatment and reduced water availability by reduced soil water recharge. Field measurements of stem sap-flow and biometry are up-scaled to stand transpiration and stand dry substance production and used to assess WUE. RUE is estimated from the ratio between the stand dry substance production and the accumulated absorbed photosynthetic active radiation over the growing season. The total stand transpiration rate for the 5 months lies between 100 and 325mm, which is fairly low compared to the Penman-Monteith transpiration for willow, reaching 400-450 mm for the same period. Mean WUE of all clones and treatments is 5.3 g/kg, which is high compared to earlier studies, while average RUE is 0.31 g/mol, which is slightly low compared to other results. Generally, all clones, except for Jorunn, seem to be better off concerning biomass production, WUE and RUE than the reference clone. Jorr, Jorunn and Loden also seem to be able to cope with the reduced water availability with increase in the water use efficiency. Tora performs significantly better than the other clones concerning both growth and efficiency in light and water use, but the effect of the dry treatment on stem growth shows sensitivity to water availability. The reduced stem growth could be due to a change in allocation patterns. (C) 2007 Elsevier Ltd. All rights reserved. (Less)

Yan-Shih Lin [et al.].- Received 16 August 2014, Accepted 16 January 2015, Published online 02 March 2015, Stomatal conductance (gs) is a key land-surface attribute as it links transpiration, the dominant component of global land evapotranspiration, and photosynthesis, the driving force of the global carbon cycle. Despite the pivotal role of gs in predictions of global water and carbon cycle changes, a global-scale database and an associated globally applicable model of gs that allow predictions of stomatal behaviour are lacking. Here, we present a database of globally distributed gs obtained in the field for a wide range of plant functional types (PFTs) and biomes. We find that stomatal behaviour differs among PFTs according to their marginal carbon cost of water use, as predicted by the theory underpinning the optimal stomatal model1 and the leaf and wood economics spectrum2, 3. We also demonstrate a global relationship with climate. These findings provide a robust theoretical framework for understanding and predicting the behaviour of gs across biomes and across PFTs that can be applied to regional, continental and global-scale modelling of ecosystem productivity, energy balance and ecohydrological processes in a future changing climate.

A new method for optimising threshold values of dry/wet spells is evaluated. A set of indices is used to find the best threshold giving good correspondence between the frequency of dry/wet spells in Hadley Centre regional model (HadRM3) output, reference observations with predetermined thresholds, and area-averaged observations. The analyses focus on selected model gridboxes in 3 different European climate regimes (Sweden, UK, Italy), where station data are available from several locations. In addition, a pan-European analysis using the European Climate Assessment (ECA) dataset is carried out. Generally, there is good agreement between point observations and the corresponding area average using the common thresholds of 0.1 or 1.0 mm with observational data. Applying the optimal thresholds on the model output is important, as it typically results in substantially better agreement between the simulated and observed series of dry/wet days. The fitted optimal pan-European dry/wet threshold is (1) 0.47 or 0.15 mm, depending on model version, for the observed point data threshold of 0.1 mm, and (2) 1.2 or 0.56 mm, depending on model version, for the threshold of 1.0 mm.

In boreal and nemoboreal forests, tree frost hardiness is modified in reaction to cues from day length and temperature. The dehardening processes in Norway spruce, Picea abies, could be estimated to start when the daily mean temperature is above 5 degreesC for 5 days. Bud burst will occur approximately after 120-170 degree-days above 5 degreesC. dependent on genetic differences among provenances. A reduced cold hardiness level during autumn and spring and an advanced onset of bud burst are expected impacts of projected future global wart-ning. The aim of this study was to test if this will increase the risk for frost damage caused by temperature backlashes. This was tested for Sweden by comparing output from the Hadley Centre regional climate model, HadRM3H, for the period 1961-1990 with future IPCC scenario SRES A2 and B2 for 2070-2099. Different indices for calculating the susceptibility to frost damage were used to assess changes in frost damage risk. The indices were based on: (1) the start of dehardening; (2) the severity of the temperature backlash: (3) the timing of bud burst: and (4) the cold hardiness level. The start of dehardening and bud burst were calculated to occur earlier all over the country. which is in line with the overall warming in both climate change scenarios. The frequency of temperature backlashes that may cause frost damage was calculated to increase in the southern part, an effect that became gradually less pronounced towards the north. The different timing of the onset of dehardening mainly caused this systematic latitudinal pattern. In the south, it occurs early in the year when the seasonal temperature progression is slow and large temperature variations occur. In the north, dehardening will occur closer to the spring equinox when the temperature progression is faster. (C) 2004 Elsevier B.V. All rights reserved. (Less)

Statistical downscaling models for precipitation in Scania, southern Sweden, have been developed and applied to calculate the changes in the future Scanian precipitation climate due to projected changes in the atmospheric composition. The models are based on multiple linear regression, linking large-scale predictors at monthly time resolution to regional statistics of daily precipitation on a monthly basis. To account for spatial precipitation variability within the area, the precipitation statistics were derived for different regions in Scania. The final downscaling models, developed for different regions and seasons, use atmospheric circulation, large-scale humidity and precipitation as predictors. Among the precipitation statistics examined, only the models for estimating the mean precipitation and the frequency of wet days were skilful. Based on the Canadian Global Circulation Model 1 (CGCM1), a future scenario of these two statistics was created. The downscaled scenario shows a significant increase of the annual mean precipitation by about 10% and a slight decrease in the frequency of wet days, indicating an increase in the precipitation amounts as well as in the precipitation intensity. The main increase of precipitation amounts and intensity occur during winter, while the summer precipitation amounts decrease slightly. The seasonal changes found in precipitation are likely attributed to changes in the westerly flow of the atmospheric circulation.

Two climate model simulations made with the Rossby Centre regional Atmospheric model version 1 (RCA1) are evaluated for the precipitation climate in Scania, southernmost Sweden. These simulations are driven by the HadCM2 and the ECHAM4/OPYC3 global circulation models (GCMs) for 10 years. Output from the global and the regional simulations are compared with an observational data set, constructed from a dense precipitation gauge network in Scania. Area-averaged time series corresponding to the size and location of the RCA1 grid points in Scania have been created (the Scanian Data Set). This data set was compared to a commonly used gridded surface climatology provided by the Climatic Research Unit (CRU). Relatively large differences were found, mainly due to the fact that the CRU-climatology uses fewer stations and lacks a correction for rain-gauge under-catch. This underlines the importance of the data set chosen for model evaluations. The validation is carried out at a large scale including the whole area of Scania and at the finest resolution of RCA1 (the grid point level). When integrated over the whole area of Scania, RCA1 improves the shape of the annual precipitation cycle and the inter-annual variability compared to output from the GCMs. The RCA1 control climate is generally too wet compared to the observations. At the grid point level, RCA1 improves the simulation of the variability compared to the GCMs. There is a strong positive correlation between precipitation and altitude in all seasons in the observations. This relationship is, however, much weaker and even reversed in the RCA1 simulations. Analysis of the dense rain gauge network reveals features of spatial variability at around 20–35 km in the area and indicates that a finer resolution is needed if the spatial variability in the area is to be better captured by RCA1.

The aim of this study is to analyse the spatial variability of meso-scale precipitation in Scania and to assess the influence of synoptic scale atmospheric circulation. The modes of spatial variation are revealed by EOF analysis of monthly precipitation totals between 1963 and 1990, which were obtained from a dense rain-gauge network in Scania. southern Sweden. The influence of local physiography on the spatial distribution of precipitation is assessed by GIS techniques using a digital elevation model of Scania. The relation to synoptic scale atmospheric circulation is analysed using regional circulation indices and weather types. It is shown that the daily precipitation distribution in the area is significantly influenced by synoptic scale pressure patterns. Nevertheless. the covariability of the monthly precipitation within Scania is high. About 80% of the precipitation variability is connected to the passage of low-pressure centres over or close to the region. which are likely to produce precipitation over the whole area. A wind-direction dependency found in the distribution indicates that there might be a limit between precipitation regimes within the landscape. Topography greatly influences the spatial distribution in Scania. The distribution of land and surrounding sea is also an important factor and makes the relationship between physiography and precipitation rather complex. The physiographical effects vary over a single year. The dampening effect of the sea on the atmospheric temperature influences the local stability in coastal areas and results in seasonally dependent precipitation patterns. (Less)

A method for calculating circulation indices and weather types following the Lamb classification is applied to southern Scandinavia. The main objective is to test the ability of the method to describe the atmospheric circulation over the area, and to evaluate the extent to which the pressure patterns determine local precipitation and temperature in Scania, southernmost Sweden. The weather type classification method works well and produces distinct groups. However, the variability within the group is large with regard to the location of the low pressure centres, which may have implications for the precipitation over the area. The anticyclonic weather type dominates, together with the cyclonic and westerly types. This deviates partly from the general picture for Sweden and may be explained by the southerly location of the study area. The cyclonic type is most frequent in spring, although cloudiness and amount of rain are lowest during this season. This could be explained by the occurrence of weaker cyclones or low air humidity during this time of year. Local temperature and precipitation were modelled by stepwise regression for each season, designating weather types as independent variables. Only the winter season-modelled temperature and precipitation show a high and robust correspondence to the observed temperature and precipitation, even though

The deposition of 137Cs over the province of Ska>ne (an area of about 100]100 km2) in the southern part of Sweden has been investigated. The origin of the deposition of 137Cs is, in about equal parts, from nuclear weapons tests and from the Chernobyl nuclear accident and amounts to about 1}3 kBq/m2. The activity concentrations of 134Cs and 137Cs in soil samples from 16 sites distributed in a grid pattern over the investigated area have been measured and the depositions from the nuclear weapons tests and from the Chernobyl accident have been separated. These pre- and post Chernobyl activities have been compared with depositions calculated from measurements of the activity concentrations of 134Cs and 137Cs in precipitation at two places and from measurements of the precipitation from a network of between 113 and 143 precipitation stations. Comparisons with in situ measurements and with aerial survey measurements have also been made. The agreement is good and it has been possible to gain a good and detailed knowledge in retrospect of the deposition from measurements of the deposition per mm of precipitation from just a few stations, and of the precipitation from a network of stations. (Less)

A warmer climate may increase the risk of attacks by insect pests on agricultural crops, and questions on how to adapt management practice have created a need for impact models. Phenological models driven by climate data can be used for assessing the potential distribution and voltinism of different insect species, but the quality of the simulations is influenced by a range of uncertainties. In this study, we model the temperature-dependent activity and development of the Colorado potato beetle, and analyse the influence of uncertainty associated with parameterization of temperature and day length response. We found that the developmental threshold has a major impact on the simulated number of generations per year. Little is known about local adaptations and individual variations, but the use of an upper and a lower developmental threshold gave an indication on the potential variation. The day length conditions triggering diapause are known only for a few populations. We used gridded observed temperature data to estimate local adaptations, hypothesizing that cold autumns can leave a footprint in the population genetics by low survival of individuals not reaching the adult stage before winter. Our study indicated that the potential selection pressure caused by climate conditions varies between European regions. Provided that there is enough genetic variation, a local adaption at the northern distribution limit would reduce the number of unsuccessful initiations and thereby increase the potential for spreading to areas currently not infested. The simulations of the impact model were highly sensitive to biases in climate model data, i.e. systematic deviations in comparison with observed weather, highlightening the need of improved performance of regional climate models. Even a moderate temperature increase could change the voltinism of Leptinotarsa decemlineata in Europe, but knowledge on agricultural practice and strategies for countermeasures is needed to evaluate changes in risk of attacks.

New tools are needed to assess carbon uptake for different forest management strategies and species composition where other methods, e.g. flux towers, fail due to large spatial integration. The aim of this study is to assess the possibility to use water use efficiency (WUE) as a means for up scaling leaf carbon uptake to forest stand scale. If the relationships hold, carbon uptake may be calculated from sap flow measurements.

Intrinsic water-use efficiency (iWUE) characterizes the physiological control on the simultaneous exchange of water and carbon dioxide in terrestrial ecosystems. Knowledge of iWUE is commonly gained from leaf-level gas exchange measurements, which are inevitably restricted in their spatial and temporal coverage. Flux measurements based on the eddy covariance (EC) technique can overcome these limitations, as they provide continuous and long-term records of carbon and water fluxes at the ecosystem scale. However, vegetation gas exchange parameters derived from EC data are subject to scale-dependent and method-specific uncertainties that compromise their ecophysiological interpretation as well as their comparability among ecosystems and across spatial scales. Here, we use estimates of canopy conductance and gross primary productivity (GPP) derived from EC data to calculate a measure of iWUE (G1 ,"stomatal slope") at the ecosystem level at six sites comprising tropical, Mediterranean, temperate, and boreal forests. We assess the following six mechanisms potentially causing discrepancies between leaf and ecosystem-level estimates of G1 : 1) non-transpirational water fluxes; 2) aerodynamic conductance; 3) meteorological deviations between measurement height and canopy surface; 4) energy balance non-closure; 5) uncertainties in NEE partitioning; and 6) physiological within-canopy gradients. Our results demonstrate that an unclosed energy balance caused the largest uncertainties, in particular if it was associated with erroneous latent heat flux estimates. The effect of aerodynamic conductance on G1 was sufficiently captured with a simple representation. G1 was found to be less sensitive to meteorological deviations between canopy surface and measurement height and, given that data are appropriately filtered, to non-transpirational water fluxes. Uncertainties in the derived GPP and physiological within-canopy gradients and their implications for parameter estimates at leaf and ecosystem level are discussed. Our results highlight the importance of adequately considering the sources of uncertainty outlined here when EC-derived WUE is interpreted in an ecophysiological context. This article is protected by copyright. All rights reserved.