Evaporative water loss of 1.42 million global lakes

Lakes, covering approximately 5 million km² of Earth's land area, are crucial components of global ecological and hydrological systems. They support biodiversity, agriculture, and human water resources. A significant amount of water is lost from lakes through evaporation, driven by the vapor pressure gradient at the water-atmosphere interface. This evaporative water loss is influenced by lake surface area and evaporation rate, both of which are geographically variable and sensitive to environmental change, including warming temperatures, increased solar radiation, and changes in lake ice cover due to climate change and drought conditions. Accurate quantification of spatiotemporal changes and the drivers of this water loss is essential for effective water resource management and adaptation strategies. However, the lack of globally consistent and locally practical datasets has previously hindered accurate global-scale quantification of evaporative water loss from lakes. Challenges like cloud contamination in satellite imagery, difficulties in quantifying lake heat storage, and the complexities of lake circulation modeling have impacted the accuracy of previous estimations. Existing global studies have primarily focused on evaporation rate changes, neglecting the overall evaporation volume. This study argues that evaporation volume, considering the dynamics of lake area and freeze-thaw cycles, provides a more comprehensive assessment of lake water balance and better informs water resource management.

Previous research on lake evaporation has faced limitations due to data scarcity and methodological challenges. While some studies have focused on evaporation rates, a comprehensive, global-scale assessment of evaporative water volume has been lacking. Existing datasets often suffer from inconsistencies in spatial and temporal resolution, hindering accurate estimations. The spatial heterogeneity of evaporation rates and water areas, influenced by factors like climate, geography, and lake morphology, further complicates the task of accurate global-scale estimations. Studies focusing solely on evaporation rates cannot fully capture the complexity of lake water balance dynamics, especially in the context of lake area dynamics and freeze/thaw cycles. Therefore, this study aimed to address these limitations by developing a novel dataset that offers a more complete picture of global lake evaporation.

This study created the first global lake evaporative volume (GLEV) dataset, encompassing monthly evaporative water loss data for 1.42 million lakes (≥10⁵ m²) from 1985 to 2018. Both natural and artificial lakes (reservoirs) were included. Monthly evaporation volume (VE) for each lake was calculated as a function of evaporation rate (Elak), lake surface area (A), and the fraction of ice duration (fLake). The monthly lake surface area was reconstructed using a Landsat-based global surface water dataset (GSWD), and the monthly fractions of ice duration were modeled using temperature and freeze/thaw lag information. Lake open water areas were calculated by multiplying lake surface area by the fraction of open water duration. Monthly data were aggregated to annual values for trend analysis. The evaporation rate (Ek) was calculated using a newly developed algorithm based on the Penman combination equation, incorporating net shortwave radiation, heat storage change, wind speed, latent heat of vaporization, and vapor pressure. The algorithm incorporated improvements to account for the heat storage capacity of deep lakes. Monthly lake surface area (Ak) was reconstructed by combining the GSWD and HydroLAKES datasets, using a weighting approach to integrate annual and monthly data. The ice duration was modeled using a developed algorithm considering temperature and a freeze/thaw lag, fitting exponential and linear relationships to freeze and thaw lag data, respectively. The algorithm was validated using simulated and in-situ data. Trend attribution analysis was performed to quantify the contributions of evaporation rate, lake surface area, and lake ice duration to the observed trends in evaporation volume. Uncertainty was quantified using standard deviation of relative values, considering the propagation of uncertainties from input variables.

The total annual evaporative water loss from global lakes (excluding the Caspian Sea) from 1985 to 2018 was estimated to be 1500 ± 150 km³, which is 15.4% higher than previous model-based estimates. The spatial distribution of evaporative volume (VE) is primarily linked to lake surface area distribution (A) but is also influenced by evaporation rate (Elak). High-latitude regions (above 46° latitude) contribute 69% to the global VE, mostly during the Northern Hemisphere summer months (June-November). The average annual open water area is about 63% of the total lake surface area, indicating significant ice cover. The relationship between VE and Elak is not linear; VE increases exponentially with Elak for lakes with Elak < 1500 mm year⁻¹, but is limited by drier climates for lakes with higher Elak values. This highlights the limitation of using evaporation rate alone as a climate change indicator. The contribution of lake evaporation to total terrestrial evapotranspiration (VE/VT) varies regionally, ranging from 0.5% in basins with high soil and vegetation evapotranspiration to over 13% in basins with large lakes. Lake evaporation volume has increased in most thermal regions over the past 34 years due to increases in both evaporation rate and lake area. The largest increase in evaporation rate is observed in the Northern Frigid region (3.7% decade⁻¹). Globally, lake evaporation volume increased at a rate of 2.1 ± 1.6% decade⁻¹ (31.2 ± 24 km³ decade⁻¹). Reservoirs contribute significantly to the total evaporation, with 6715 reservoirs accounting for 16% of global evaporation volume (235 km³ year⁻¹) despite only representing 5% of the total lake storage capacity. Evaporative water loss from reservoirs has increased at a rate of 5.4% decade⁻¹. Trend attribution analysis indicates that increasing evaporation rate (58%), decreasing lake ice coverage (23%), and increasing lake surface area (19%) contributed to the observed increase in global lake evaporation volume. For many lakes, evaporation variability is dominated by non-precipitation forcings like temperature changes, while for others, it's controlled by hydrological conditions and reservoir operations that affect lake surface area.

The findings underscore the accelerating global evaporative water loss under climate change, particularly in high-latitude and high-altitude regions. Increased temperatures in these regions lead to greater heat uptake by lakes (due to lower albedo of liquid water compared to ice) and larger open water areas for evaporation. The resulting increased evaporation can alter local and regional hydroclimatic processes, requiring more accurate representation in Earth System Models. The substantial evaporation from reservoirs, exceeding the global trend, adds stress to water resources, especially considering growing demands from agriculture, industry, and domestic use. The GLEV dataset, providing long-term monthly evaporation data for 1.42 million lakes, offers a valuable resource for researchers and decision-makers, enabling improved water availability estimations, especially during droughts, and improved simulations of moisture transport. The study contributes to the growing understanding of global water bodies and their response to climate change.

This study presents the first comprehensive global dataset (GLEV) of lake evaporative water loss, revealing a significant and increasing contribution of lake evaporation to the global water cycle. The increasing evaporation volume is driven by a combination of rising evaporation rates and changes in lake surface area due to factors such as reduced ice cover and reservoir expansion. The results highlight the critical need for accurate accounting of lake evaporation in water resource management and climate modeling. Future research could focus on improving the accuracy of evaporation rate estimations and incorporating more detailed information on lake bathymetry and water quality to refine the model and expand the analysis to a wider range of lake characteristics.

While this study provides a significant advancement in understanding global lake evaporation, several limitations exist. The accuracy of the evaporation rate estimations depends on the accuracy of the meteorological input data, and uncertainties in these data might propagate into the evaporation volume estimates. The modeling of ice duration relies on simplified assumptions about lake ice phenology, potentially impacting the accuracy of estimates, particularly for large lakes with complex ice cover dynamics. The study also doesn't fully account for the effects of groundwater interactions on lake water balance. Future research could address these limitations by integrating higher-resolution meteorological data and more sophisticated ice phenology models.