Appraisal of Budyko formula in calculating long‐term water balance in humid watersheds of southern China

The Budyko formula for estimating the long‐term average annual evaporation is applied to calculate the long‐term water balance in 29 humid watersheds of southern China. As a result of overestimation of evaporation, the long‐term average annual runoff is underestimated, with the Nash‐Sutcliffe efficiency (NSE) at just ? 17%. A one‐variable linear regression model is employed to find that the Budyko scatter and the relative errors of Budyko runoff and evaporation estimates are all closely related to the long‐term aridity index. Through combining the original Budyko formula with the different linear regression models for estimating the Budyko estimation errors, three forms of revised Budyko equation for estimating the long‐term average annual runoff are derived, with all their NSE values to be around 66%. After calibration, both one‐parameter Turc‐Pike and one‐parameter Fu equations lead to the NSE value of 60% in estimating long‐term average annual runoff. Two conclusions are made, with the first one being that, the nonparametric Budyko formula, although very intuitive and very simple, does not apply well in calculating long‐term water balance in 29 humid watersheds in southern China. The second one is that, the parametric evaporation formulae, with locally optimized parameter values, can achieve better accuracy in estimating long‐term average annual evaporation and runoff than the nonparametric Budyko evaporation formula. Copyright © 2011 John Wiley & Sons, Ltd.     

Long‐term investigations of the snow cover in a subalpine semi‐forested catchment

To improve spring runoff forecasts from subalpine catchments, detailed spatial simulations of the snow cover in this landscape is obligatory. For more than 30 years, the Swiss Federal Research Institute WSL has been conducting extensive snow cover observations in the subalpine watershed Alptal (central Switzerland). This paper summarizes the conclusions from past snow studies in the Alptal valley and presents an analysis of 14 snow courses located at different exposures and altitudes, partly in open areas and partly in forest. The long‐term performance of a physically based numerical snow–vegetation–atmosphere model (COUP) was tested with these snow‐course measurements. One single parameter set with meteorological input variables corrected to the prevailing local conditions resulted in a convincing snow water equivalent (SWE) simulation at most sites and for various winters with a wide range of snow conditions. The snow interception approach used in this study was able to explain the forest effect on the SWE as observed on paired snow courses. Finally, we demonstrated for a meadow and a forest site that a successful simulation of the snowpack yields appropriate melt rates. Copyright © 2006 John Wiley & Sons, Ltd.     

Long‐term change of daily and multi‐daily precipitation in southern Sweden

Daily rain series from southern Sweden with records dating back to the 1870s have been analysed to investigate the trends of daily and multi‐day precipitation of different return periods with emphasis on the extremes. Probabilities of extreme storms were determined as continuously changing values based on 25 years of data. An extra set of data was used to investigate changes in Skåne, the southernmost peninsula of Sweden. Another 30‐year data set of more than 200 stations of a dense gauge network in Skåne was used to investigate the relation between very large daily rainfall and annual precipitation. The annual precipitation has increased significantly all over southern Sweden due to increased winter precipitation. There is a trend of increasing maximum annual daily precipitation at only one station, where the annual maximum often occurs in winter. The number of events with a short return period is increasing, but the number of more extreme events has not increased. Daily and multi‐daily design storms of long return periods determined from extreme value analysis with updating year by year are not higher today than during the last 100 years. The largest daily storms are not related to stations with annual rainfall but seem to occur randomly. Copyright © 2013 John Wiley & Sons, Ltd.     

Erodibility of low‐compaction steep‐sloped reclaimed surface mine lands in the southern Appalachian region,USA

The use of loose spoils on steep slopes for surface coal mining reclamation sites has been promoted by the US Department of Interior, Office of Surface Mining for the establishment of native forest, as prescribed by the Forest Reclamation Approach (FRA). Although low‐compaction spoils improve tree survival and growth, erodibility on steep slopes was suspected to increase. This study quantified a combined K_C factor (combining the effects of the soil erodibility K factor and cover management C) for low compaction, steep‐sloped (>20°) reclaimed mine lands in the Appalachian region, USA. The combined K_C factor was used because standard Unit Plot conditions required to separate these factors, per Revised Universal Soil Loss Equation (RUSLE) experimental protocols, were not followed explicitly. Three active coal mining sites in the Appalachian region of East Tennessee, each containing four replicate field plots, were monitored for rainfall and sediment yields during a 14‐month period beginning June 2009. Average cumulative erosivity for the study sites during the monitoring period was measured as 5248.9 MJ·mm·ha^?1·h^?1. The K_C ranged between 0.001 and 0.05 t·ha·h·ha^?1·MJ^?1·mm^?1, with the highest values occurring immediately following reclamation site construction as rills developed (June – August 2009). The K_C for two study sites with about an 18–20 mm spoil D₈₄ were above 0.01 t·ha·h·ha^?1·MJ^?1·mm^?1 during rill development, and below 0.003 t·ha·h·ha^?1·MJ^?1·mm^?1 after August 2009 for the post‐rill development period. The K_C values for one site with a 40 mm spoil D₈₄ were never above 0.008 t·ha·h·ha^?1·MJ^?1·mm^?1 and also on average were lower, being more similar to the other two sites after the rill development period. Based on an initial K_C factor (K_e) measured during the first few storm events, the average C factor (C_e) was estimated as 0.58 for the rill development period and 0.13 for the post‐rill development period. It appears that larger size fractions of spoils influence K_C and C_e factors on low‐compaction steep slopes reclamation sites. Copyright © 2013 John Wiley & Sons, Ltd.     

Erosion of a cyclic saltmarsh in Morecambe Bay,North-West England

This paper presents results of investigations (1983–1992) into rates of change, morphology and processes occurring during the current erosional phase in a Morecambe Bay cyclic saltmarsh, in which it has narrowed from c. 1000 m (1975) to c. 150 m (1992). Monthly monitoring of marsh edge erosion and creek knickpoint retreat has revealed temporal and spatial variations. Highest frequency changes of low magnitude coincided with non-storm conditions and overmarsh tides above 5·80 m OD, which submerged the whole marsh. Less frequent changes of greater magnitude were associated with both overmarsh tides and strong onshore winds over 15 ms^?1, which generated high energy waves. The lowest frequency change of greatest magnitude occurred during an extreme onshore storm event and surge. Morphologically, during the erosional phase, a low angled landward slope was generated as erosion of the c. 0·5 m high active seaward cliff coincided with vertical accretion of 0·07 ma^?1 of relatively coarse sediment on the marsh surface immediately landward. Tidal hydrodynamics strongly influence the saltmarsh, which is confined to the upper 2·5 m of the macrotidal range (maximum c. 10·5 m). During overmarsh spring tides (maximum creek flood flow rate 0·13 ms^?1, up to bankfull level), flooding begins over lower landward creek banks before submerging the higher marsh edge. During ebb tides, water trapped by this higher edge can escape seaward only via the creeks (maximum ebb velocities 2·07 ms^?1 below bankfull level). Wave erosion also is limited to spring tides. Monthly mapping of the Kent Estuary channel pattern seaward of the saltmarsh showed that medium term higher erosion rates were related to the presence of a large channel, which lowered the adjacent creek base level and allowed larger waves to attack the marsh edge than when a sandbank flanked the marsh. Major River Kent channel shifts appear to initiate accretional or erosional phases of cyclic saltmarsh development.     

Evaluating restoration of man‐made slopes: a threshold approach balancing vegetation and rill erosion

The management of reclaimed slopes derived from industrial and civil activities (e.g. surface mining and road construction) requires the development of practical stability analysis approaches that integrate the processes and mechanisms that rule the dynamics of these ubiquitous emerging ecosystems. This work describes a new modelling approach focused on stability analysis of water‐limited reclaimed slopes, where interactive relationships between rill erosion and vegetation regulate ecosystem stability. Our framework reproduces two main groups of possible trends along the temporal evolution of reclaimed slopes: successful trends, characterized by widespread vegetation development and the effective control of rill erosion processes; and gullying trends, characterized by the progressive loss of vegetation and a sharp logistic increase in erosion rates. Furthermore, this analytical approach allows the determination of threshold values for the state variables (i.e. vegetation cover and rill erosion) that drive the system's stability, facilitating the identification of critical situations that require specific human intervention (e.g. revegetation or, in very problematic cases, revegetation combined with rill network destruction) to ensure the long‐term sustainability of the restored ecosystem. The application of our threshold analysis framework in Mediterranean‐dry reclaimed slopes derived from surface coal mining (the Teruel coalfield in central‐eastern Spain) showed a good field‐based performance. Therefore, we believe that this model is a valuable contribution for the management of water‐limited reclaimed systems, including those associated with rill erosion, as it provides a tool for the evaluation of restoration success and can play an important role in decision‐making during ecosystem restoration in severely disturbed landscapes. Copyright © 2011 John Wiley & Sons, Ltd.     

Soil water content in southern England derived from a cosmic‐ray soil moisture observing system – COSMOS‐UK

Cosmic‐ray soil moisture sensors have the advantage of a large measurement footprint (approximately 700 m in diameter) and are able to operate continuously to provide area‐averaged near‐surface (top 10–20 cm) volumetric soil moisture content at the field scale. This paper presents the application of this technique at four sites in southern England over almost 3 years. Results show the soil moisture response to contrasting climatic conditions during 2011–2014 and are the first such field‐scale measurements made in the UK. These four sites are prototype stations for a UK COsmic‐ray Soil Moisture Observing System, and particular consideration is given to sensor operating conditions in the UK. Comparison of these soil water content observations with the Joint UK Land Environment Simulator 10‐cm soil moisture layer shows that these data can be used to test and diagnose model performance and indicate the potential for assimilation of these data into hydro‐meteorological models. The application of these large‐area soil water content measurements to evaluate remotely sensed soil moisture products is also demonstrated. Numerous applications and the future development of a national COsmic‐ray Soil Moisture Observing System network are discussed. Copyright © 2016 John Wiley & Sons, Ltd.     

Rainfall,run‐off,and sediment transport dynamics in a humid mountain badland area: Long‐term results from a small catchment

In this study, we investigated rainfall, run‐off, and sediment transport dynamics (414 run‐off events and 231 events with sediment information) of a humid mountain badland area—the Araguás catchment (Central Pyrenees, Spain)—from October 2005 to September 2016. Use of this long‐term database allows characterization of the hydrological response, which consist of low‐magnitude/high‐frequency events and high‐magnitude/low‐frequency events, and identification of seasonal dynamics and rainfall‐run‐off thresholds. Our results indicate that the Araguás catchment, similarly to other humid badlands, had high hydrological responsiveness (mean annual run‐off coefficient: 0.52), a non‐linear relationship of rainfall with run‐off (common in Mediterranean environments), and seasonal hydrological and sedimentological dynamics. We created and validated a multivariate regression model to characterize the hydrological variables (stormflow and peak discharge) and sedimentological variables (mean and maximum suspended sediment concentrations and total suspended sediment load). In summer and at the beginning of autumn, the response was mainly related to rainfall intensity, suggesting a predomination of Hortonian flows. In contrast, in spring and winter, the responses were mainly related to the antecedent conditions (previous rainfall and baseflow), suggesting the occurrence of saturated excess flow processes, and the contribution of neighbouring vegetated areas. The multivariate analysis also showed that total sediment load is better predicted by a multivariate regression model that integrates pre‐event, rainfall, and run‐off variables. In general, our models provided more accurate predictions of small‐magnitude/high‐frequency events than high‐magnitude/low‐frequency events. This study highlights the high inter‐ and intra‐annual variability response in humid badland areas and that long‐term records are needed to reduce the uncertainty of hydrological and sedimentological responses in Mediterranean badland areas.     

Quantification of bank erosion in a drained agricultural lowland catchment

The long‐term and current volumes of sediment exported from stream banks were calculated as potential sources of sediment in a large pond located at the catchment outlet of a small agricultural lowland basin strongly affected by anthropogenic pressure in France. Bank erosion was measured over a short period using a network of erosion pins along a small stream (1400 m long) to quantify the material exported during a single winter (2012–2013). The material exported by this same stream over the last 69 years was quantified using an original approach involving the comparison of a compilation of three‐dimensional historical stream redesign plans that date back to 1944 with the state of the banks in 2013 (differential global positioning system and LiDAR data). The results suggest that a global trend of material loss along the stream banks monitored by erosion pins, with an average erosion rate of 17.7 mm year^?1 and an average volume of exported material of 75 t km^?1. Over 69 years, this same stream exported an average of 36 t km^?1 year^?1, and the average loss of material from the banks throughout the whole catchment was estimated to be 14 t km^?1 year^?1. The contribution of bank material to the filling of the pond over the last 10 years is between 46% and 52% based on an extrapolation of erosion pin dynamics or between 27% and 30% based on the comparison of LiDAR data to the average historical profile extrapolated for the catchment. These results suggest that bank erosion represents a major source of sediment in degraded waters in traditionally understudied agricultural lowland catchments, where anthropogenic pressures are high.     

Sediment transport measurements with tracers in very low‐energy beaches

This study investigates sediment transport at a very low‐energy backbarrier beach in southern Portugal, from a spring‐to‐neap tide period, during fair‐weather conditions. Rates and directions of transport were determined based on the application of fluorescent tracer techniques. Wind and currents were collected locally, whereas the dominant small and short‐period wind waves were characterized using a morphodynamic modelling system coupling a circulation model, a spectral wave model, and a bottom evolution model, well validated over the study area. For the recorded conditions sediment transport was small and ebb oriented, with daily transport rates below 0.02 m³ day^‐1. Tidal currents (mainly ebb velocities) were found to be the main causative forcing controlling sediment displacements. Transport rates were higher during spring tides, tending towards very small values at neap tides. Results herein reported points towards the distinction between tracer advection and tracer dispersion in this type of environment. Transport by advection was low as a consequence of the prevailing hydrodynamic conditions (H_s < 0.1 m, and max. current velocity of 0.5 m s^‐1) and the tracer adjustment to the transport layer, whereas dispersion was relatively high (few metres per day). Tracer techniques allowed distinguishing the broad picture of transport, but revealed the need for refinement in this type of environments (bi‐directional forcing by ebb and flood cycles). Copyright © 2012 John Wiley & Sons, Ltd.     

Seismic energy dissipation in deteriorating systems through low‐cycle fatigue

Energy dissipation characteristics of structural members which exhibit both strength and stiffness deterioration under imposed displacement reversals are investigated. In the experimental part, 17 reinforced concrete beam specimens were tested under constant and variable amplitude inelastic displacement cycles. The constant‐amplitude tests were employed to determine the low‐cycle fatigue behaviour of specimens where the imposed displacement amplitude was the major variable. A two‐parameter fatigue model was developed in order to express the variation of dissipated energy with the number of displacement cycles. This model was then used to predict the energy dissipation of test specimens subjected to variable‐amplitude displacement cycles simulating severe seismic excitations. It has been demonstrated that the remaining energy dissipation capacity in a forthcoming displacement cycle is dependent on the energy dissipated along the completed displacement path. Moreover, it is observed that total energy dissipation is dependent on the length of the displacement path. Copyright © 2003 John Wiley & Sons, Ltd.     

Using structure‐from‐motion to create glacier DEMs and orthoimagery from historical terrestrial and oblique aerial imagery

Increased resolution and availability of remote sensing products, and advancements in small‐scale aerial drone systems, allows observations of glacial changes at unprecedented levels of detail. Software developments, such as structure‐from‐motion (SfM), now allow users an easy and efficient method to generate three‐dimensional (3D) models and orthoimages from aerial or terrestrial datasets. While these advancements show promise for current and future glacier monitoring, many regions still suffer a lack of observations from earlier time periods. We report on the use of SfM to extract spatial information from various historic imagery sources. We focus on three geographic regions, the European Alps, high Arctic Norway and the Nepal Himalayas. We used terrestrial field photographs from 1896, high oblique aerial photographs from 1936 and aerial handheld photographs from 1978 to generate digital elevation models (DEMs) and orthophotos of the Rhone glacier, Brøggerhalvøya and the lower Khumbu glacier, respectively. Our analysis shows that applying SfM to historic imagery can generate high quality models using only ground control points. Limited camera/orientation information was largely reproduced using self‐calibrated model data. Using these data, we calculated mean ground sampling distances across each site which demonstrates the high potential resolution of resulting models. Vertical errors for our models are ±5.4 m, ±5.2 m and ±3.3 m. Differencing shows similar patterns of thinning at lower Rhone (European Alps) and Brøggerhalvøya (Norway) glaciers, which have mean thinning rates of 0.31 m a^?1 (1896–2010) to 0.86 m a^?1 (1936–2010) respectively. On these clean ice glaciers thinning is highest in the terminus region and decreasing up‐glacier. In contrast to these glaciers, uneven topography, exposed ice‐cliffs and debris cover on the Khumbu glacier create a highly variable spatial distribution of thinning. The mean thinning rate for the Khumbu study area was found to be 0.54 ± 0.9 m a^?1 (1978–2015). Copyright © 2017 John Wiley & Sons, Ltd.     

Complexity‐reduction modelling for assessing the macro‐scale patterns of historical soil moisture in the Euro‐Mediterranean region

Complexity‐reduction modelling can be useful for increasing the understanding of how the climate affects basin soil moisture response upon historical times not covered by detailed hydrological data. For this purpose, here is presented and assessed an empirical regression‐based model, the European Soil Moisture Empirical Downscaling (ESMED), in which different climatic variables, easily available on the web, are addressed for simplifying the inherent complexity in the long‐time studies. To accommodate this simplification, the Palmer Drought Severity Index, the precipitation, the elevation and the geographical location were used as input data in the ESMED model for predicting annual soil moisture budget. The test area was a large region including central Europe and Mediterranean countries, and the spatial resolution was initially set at 50 km. ESMED model calibration was made according to the soil moisture values retrieved from the Terrestrial Water Budget Data archive by selecting randomly 285 grid points (out of 2606). Once parameterized, ESMED model was performed at validation stage both spatially and temporally. The spatial validation was made for the grid points not selected in the calibration stage while the comparison with the soil moisture outputs of the Global Land Data Assimilation System–NOAH10 simulations upon the period 1950–2010 was carried out for the temporal validation. Moreover, ESMED results were found to be in good agreement with a root‐zone soil moisture product obtained from active and passive microwave sensors from various satellite missions. ESMED model was thus found to be reliable for both the temporal and spatial validations and, hence, it might represent a useful tool to characterize the long‐term dynamics of soil moisture–weather interaction. Copyright © 2013 John Wiley & Sons, Ltd.     

Spatially distributed long‐term hydrologic simulation using a continuous SCS CN method‐based hybrid hydrologic model

A continuous Soil Conservation Service (SCS) curve number (CN) method that considers time‐varied SCS CN values was developed based on the original SCS CN method with a revised soil moisture accounting approach to estimate run‐off depth for long‐term discontinuous storm events. The method was applied to spatially distributed long‐term hydrologic simulation of rainfall‐run‐off flow with an underlying assumption for its spatial variability using a geographic information systems‐based spatially distributed Clark's unit hydrograph method (Distributed‐Clark; hybrid hydrologic model), which is a simple few parameter run‐off routing method for input of spatiotemporally varied run‐off depth, incorporating conditional unit hydrograph adoption for different run‐off precipitation depth‐based direct run‐off flow convolution. Case studies of spatially distributed long‐term (total of 6 years) hydrologic simulation for four river basins using daily NEXRAD quantitative precipitation estimations demonstrate overall performances of Nash–Sutcliffe efficiency (E_NS) 0.62, coefficient of determination (R²) 0.64, and percent bias 0.33% in direct run‐off and E_NS 0.71, R² 0.72, and percent bias 0.15% in total streamflow for model result comparison against observed streamflow. These results show better fit (improvement in E_NS of 42.0% and R² of 33.3% for total streamflow) than the same model using spatially averaged gauged rainfall. Incorporation of logic for conditional initial abstraction in a continuous SCS CN method, which can accommodate initial run‐off loss amounts based on previous rainfall, slightly enhances model simulation performance; both E_NS and R² increased by 1.4% for total streamflow in a 4‐year calibration period. A continuous SCS CN method‐based hybrid hydrologic model presented in this study is, therefore, potentially significant to improved implementation of long‐term hydrologic applications for spatially distributed rainfall‐run‐off generation and routing, as a relatively simple hydrologic modelling approach for the use of more reliable gridded types of quantitative precipitation estimations.     

Increasing water losses from snow captured in the canopy of boreal forests: A case study using a 30 year data set

Water losses from snow intercepted by forest canopy can significantly influence the hydrological cycle in seasonally snow‐covered regions, yet how snow interception losses (SIL) are influenced by a changing climate are poorly understood. In this study, we used a unique 30 year record (1986–2015) of snow accumulation and snow water equivalent measurements in a mature mixed coniferous (Picea abies and Pinus sylvestris ) forest stand and an adjacent open area to assess how changes in weather conditions influence SIL. Given little change in canopy cover during this study, the 20% increase in SIL was likely the result of changes in winter weather conditions. However, there was no significant change in average wintertime precipitation and temperature during the study period. Instead, mean monthly temperature values increased during the early winter months (i.e., November and December), whereas there was a significant decrease in precipitation in March. We also assessed how daily variation in meteorological variables influenced SIL and found that about 50% of the variation in SIL was correlated to the amount of precipitation that occurred when temperatures were lower than ?3 °C and to the proportion of days with mean daily temperatures higher than +0.4 °C. Taken together, this study highlights the importance of understanding the appropriate time scale and thresholds in which weather conditions influence SIL in order to better predict how projected climate change will influence snow accumulation and hydrology in boreal forests in the future.     

Long‐term discharge prediction for the Turnu Severin station (the Danube) using a linear autoregressive model

The long‐term trends of yearly discharge time series and runoff variability at seven stations along the River Danube are identified. The results of statistical analysis of discharge time series indicate the period around the year 1860 was the driest decade in central and eastern Europe since 1840. In these years, the mean annual air temperature in central Europe was lower by about 1 °C compared with the 1990s. It is important to notice that the two driest decades (around 1860s and 1990s) of the instrumental era occurred in very different temperature conditions. The 28–31 years; 20–21 years; 14 years, as well as 4·2, 3·6, and 2·4 years fluctuations of annual discharge in the River Danube were found. Also, the long‐term streamflow prediction based on stochastic modelling methods is treated. Harmonic models and the Box–Jenkins methods were used. The predictions of yearly River Danube discharge time series were made for two decades ahead. From the stochastic models it follows that the annual discharge in the Danube at Turnu Severin station should reach its local maximum within the years 2004–06. The period 2015–19 should be dry. Copyright © 2005 John Wiley & Sons, Ltd.     

Vegetation controls on small‐scale runoff and erosion dynamics in a degrading dryland environment

This paper investigates the controls of vegetation on runoff and erosion dynamics in the dryland environment of Jornada, New Mexico, USA. As the American southwest has seen significant shifts in the dominant vegetation species in the past 150 years, an understanding of the vegetation effects on hydrological and erosional processes is vital for understanding and managing environmental change. Small‐scale rainfall simulations were carried out to identify the hydrological and erosional processes resulting from the grassland and shrubland vegetation species. Results obtained using tree‐regression analysis suggested that the primary vegetation control on runoff and erosion is the shrub type and canopy density, which directly affects the local microtopographic gradient of mounds beneath the shrubs. Significant interactions and feedbacks were found to occur among the local mound gradient, crust cover, soil aggregate stability and antecedent soil moisture between the different vegetation species for both the runoff and erosion responses. Although some of the shrub species were found to produce higher sediment yields than the grass species, the distinguishing feature of the grassland was the significantly higher enrichment in the fine sediment fraction compared to all other surface cover types. This enrichment in fines has important implications for nutrient movement in such environments. Copyright © 2009 John Wiley & Sons, Ltd.     

Thermal effects of lateral supra‐permafrost water flow around a thermokarst lake on the Qinghai–Tibet Plateau

Both the inflow and outflow of supra‐permafrost water to lakes play important roles in the hydrologic process of thermokarst lakes. The accompanying thermal effects on the adjacent permafrost are required for assessing their influences on the development of thermokarst lakes. For these purposes, the lake water level, temperature dynamics, and supra‐permafrost water flow of a lake were monitored on the Qinghai‐Tibet Plateau. In addition, the spatial and temporal variation of the active layer thickness and permafrost distribution around the lake were investigated by combining ground penetrating radar, electrical resistivity tomography, and borehole temperature monitoring. The results revealed that the yearly unfrozen supra‐permafrost water flow around the lake lasted approximately 5 months. The temperature and water level measurements during this period indicate that the lake water was recharged by relatively colder supra‐permafrost water from the north‐western lakeshore and was discharged through the eastern lakeshore. This process, accompanied by heat exchange with the underlying permafrost, might cause a directional difference of the active layer thickness and permafrost characteristics around the lake. Specifically, the active layer thickness variation was minimal, and the ice‐rich permafrost was found adjacent to the lakeshore along the recharge groundwater pathways, whereas a deeper active layer and ice‐poor permafrost were observed close to the lakeshore from which the warm lake water was discharged. This study suggests that the lateral flow of warm lake water can be a major driver for the rapid expansion of thermokarst lakes and provides clues for evaluating the relationships between the thermokarst expansion process and climate warming.