Multi‐year observations of the high mountain water cycle in the Langtang catchment, Central Himalaya

Abstract The Langtang catchment is a high mountain, third order catchment in the Gandaki basin in the Central Himalaya (28.2°N, 85.5°E), that eventually drains into the Ganges. The catchment spans an elevation range from 1400 to 7234 m a.s.l. and approximately one quarter of the area is glacierized. Numerous research projects have been conducted in the valley during the last four decades, with a strong focus on the cryospheric components of the catchment water balance. Since 2012 multiple weather stations and discharge stations provide measurements of atmospheric and hydrologic variables. Full weather stations are used to monitor at an hourly resolution all four radiation components (incoming and outgoing shortwave and longwave radiation; SWin/out and LWin/out), air temperature, humidity, wind speed and direction, and precipitation, and cover an elevational range of 3862–5330 m a.s.l. Air temperature and precipitation are monitored along elevation gradients for investigations of the spatial variability of the high mountain meteorology. Dedicated point‐scale observations of snow cover, depth and water equivalent as well as ice loss have been carried out over multiple years and complement the observations of the water cycle. All data presented is openly available in a database and will be updated annually.

have been conducted in the valley during the last four decades, with a strong focus on the cryospheric components of the catchment water balance. Since 2012 multiple weather stations and discharge stations provide measurements of atmospheric and hydrologic variables. Full weather stations are used to monitor at an hourly resolution all four radiation components (incoming and outgoing shortwave and longwave radiation; SW in/out and LW in/out ), air temperature, humidity, wind speed and direction, and precipitation, and cover an elevational range of 3862-5330 m a.s.l. Air temperature and precipitation are monitored along elevation gradients for investigations of the spatial variability of the high mountain meteorology. Dedicated point-scale observations of snow cover, depth and water equivalent as well as ice loss have been carried out over multiple years and complement the observations of the water cycle. All data presented is openly available in a database and will be updated annually.

| DATA SET NAME
The Langtang (Central Himalaya) meteorological and streamflow dataset.

| SITE DESCRIPTION
The Langtang catchment, located in the Central Himalaya in Nepal (28.2 N, 85.5 E), covers an area of approximately 585 km 2 and spans an elevation range from 1400 m a.s.l. at the confluence of the Langtang River with the main Trisuli River to 7234 m a.s.l. at the peak of Langtang Lirung. Approximately 25% of the area is glacierized, with extensive debris cover on glacier termini below 5200 m. The climate is dominated by the monsoon, and characterized by synoptic scale easterly flow in summer (June to September) and westerly flow between October and May, with near surface wind directions modulated by valley circulation. Between 70% and 90% of the total annual precipitation falls during monsoon (Immerzeel et al., 2014) but total precipitation varies from around 2000 mm year À1 at 2300 m to less than 1000 mm year À1 14 km further to the east at 3900 m (Immerzeel et al., 2014). The catchment entrance at Syafru Besi (1400 m) is accessible by road, while the rest of the catchment can only be reached on foot or by helicopter. The highest permanent settlement, Kyanjing, is located at 3900 m, close to the main automatic weather station (AWS) and discharge measurement site.

| INSTRUMENTATION
Data from all stations is downloaded manually once or twice a year on site and uploaded to the database annually, and stations are serviced in the process. Data collected during sensor failure or when a station was not upright or otherwise damaged are set to NA. Stations repeatedly failed either due to battery malfunction, damage due to an earthquake or avalanches as well as high flow events. All data is stored in the database as collected and no processing is performed, with the exception of discharge.

| Meteorology
Research in the catchment, mainly of the cryosphere, dates back to the 1980s (see Higuchi (1993) for an overview). In 1987 the Govern-  Table 2), measuring all four radiation components (SW in/out , LW in/out ), wind speed and direction, air temperature and relative humidity as well as precipitation and snow depth. Snow depth at all stations in the catchment is stored as distance from sensor to surface and actual snow depth has to be computed by the user. An additional AWS with the same sensors as above, excluding precipitation measurements is operational on Yala Glacier at approximately 5330 m a.s.l. This station is reinstalled as needed due to the wasting ice. All AWS sample data at 10 min interval, which are subsequently aggregated to hourly intervals. Small weather stations (MicroMet; sensor type: Lufft WS500-UMB, Fellbach, Germany) that measure wind speed (accuracy: ±0.3 m/s or ±3%) and direction (<3 at wind speed >1 m/s), air temperature (±0.2 C) and relative humidity (±2%) are located on debris-covered ice (Lirung Glacier, 4200 m a.s.l.) and a moraine (Morimoto, 4919 m a.s.l.). In addition a transect of MicroMet stations (sensor type for wind: Lufft WS300-UMB, Fellbach, Germany; sensor type for temperature and humidity: Campbell Scientific 215, Logan, USA) measuring wind speed and direction (accuracies as above) as well as air temperature (±0.3 C (at 25 C) ±0.4 C (5-40 C)) and relative humidity (±4% at 25 C) was placed on the Yala Plateau, (4800, 5278 and 5504 m a.s.l.) between October 2016 and April 2017 , not shown in Figure 1. All MicroMet stations sample at 15 min intervals and data is published in the same format.
For 2 weeks on Lirung Glacier independent station) and multiple weeks on Yala Glacier  mounted on the Yala Glacier AWS mentioned above) eddy covariance was measured to determine turbulent exchanges over the glacier surface, using the Campbell Scientific IRGASON sensor (Logan, USA).
Data is collected at 10 Hz interval and stored at 5 and 60 min interval for Lirung and Yala respectively. On Lirung Glacier, low frequency radiation (CNR1), air temperature and relative humidity (HC2S3) is also available, sampled and stored in 10 min intervals.
A network of sensors was installed to observe spatiotemporal variability in precipitation. Pluviometers at 4919 m a.s.l. (Morimoto, Table 2) on the northern slope and 4452 m a.s.l. (Langshisha) on the southern slope measured total precipitation, snow height and wind speed, temperature and relative humidity (Table 1) Table 2). These sensors store the time step at each individual tip, which then needs to be aggregated by the user to a final time interval and associated precipitation volume. Precipitation data is not corrected for undercatch. However wind speed is recorded alongside pluviometers which allows for simple corrections that are especially relevant during snowfall (Thériault et al., 2012).

| Hydrology
Stage height was measured on the main stem of Langtang River at 3850 m a.s.l. using a radar level sensor (Table 2,

| RESEARCH INSIGHTS
Discharge measured at the main gauging station varies between below 5 m 3 s À1 (winter) and 20 m 3 s À1 (monsoon, Figure 2). A steep increase in discharge occurs in May, when snow melt at high altitudes coincides with the first monsoon rains in lower regions, soil thaws and becomes saturated (Figure 2). Variability in soil moisture after saturation is driven by precipitation events that result in high altitude snow fall and almost immediate melt shortly thereafter. By early June, the soil is near-saturated due to the snow melt, monsoon precipitation begins, and high elevation snow and ice melt is occurring. As a result T A B L E 1 Sensor specifications for the AWS and pluviometers Note: Sampling interval at all main stations is 10 min, data is available at hourly intervals unless otherwise stated in the metadata. CS stands for the manufacturer Campbell Scientific (Logan, USA). Location of manufacturer are given in italics after the sensor type. Sensor accuracies are shown in brackets and with the metadata of each station. Previous research in the catchment shows a strong elevational gradient of precipitation, which peaks above 1500 mm annually around 3000 m a.s.l. (Collier & Immerzeel, 2015) and rapidly decreases up-and down-valley, with station data showing a drop from 1819 mm at 2370 m a.s.l. to just 867 mm at 3857 m a.s.l. (Immerzeel et al., 2014). At higher elevations the valley floors are generally drier than the southern and northern slopes that receive more orographic precipitation (Collier & Immerzeel, 2015

DATA AVAILABILITY STATEMENT
Data from the AWS and precipitation measurements as well as all discharge data are available and updated via the regional database of ICIMOD (full URL for all combined data: http://rds.icimod.org/ Home/Data?any=Langtang). To access data in the database, an initial registration is necessary, which enables immediate access. Each individual dataset includes a brief description of the station, while metadata describing sensor specifications as well as details on the derivation of discharge data including original stage height measurements is supplied with the data file upon download. Each individual dataset has a separate DOI provided in the respective metadata. Raw data or any additional details on sensor setups and site conditions for all stations may be acquired via the authors.
F I G U R E 2 Discharge (Q), snow depth (SD), soil moisture (SM), precipitation (P) and air temperature (T air ) data from 21 April to 14 July 2018. The dashed red line marks the 0 C line. Note that discharge is measured at 3850 m a.s.l., soil moisture at 4807 m a.s.l. and all other variables at 5100 m a.s.l.