Weather stations on tropical glaciers: A multivariate dataset for cryospheric and climate research of Peruvian glaciers

Wilson Suarez^1*Cristobal Lucero¹Mónica Villacorta²

• ¹Servicio Nacional de Meteorología e Hidrología del Perú (SENAMHI), Lima, Peru
• ²Universidad Nacional Agraria La Molina, Lima, Peru

Tropical glaciers serve as crucial sources of fresh water, particularly in arid tropical regions, where glacial runoff supports human consumption and economic activities such as agriculture, livestock farming, and tourism. Glaciers, due to their rapid response to melting conditions, are widely regarded key indicators of climate change (Kaser and Osmaton, 2002), in recent years a significant loss of glacier mass was reported along with changes in the climate of periglacial areas (ICCI, 2023), it was also estimated the tropical Andean glaciers region have lost around 42% of its total area with significant losses in regions below 5000 masl in the 1990 -2020 period (Turpo-Cayo et al., 2022), further alarming authorities and researchers due to its already discussed repercussions on other environmental and economic elements (Bradley et al., 2006;Kaser et al., 2005;Vergara et al., 2007).In South America, the extent of glaciers covers several countries from Bolivia to Venezuela, with Peru having a large number of glaciers in its 18 glacial mountain ranges located mainly between the center and south regions, according to recent measurements it has a total glacial extension of around 1050.32 km 2 registered in 2020 thus marking a loss of 6.5 % since 2017 (INAIGEM, 2023). Studies on the future of these glaciers indicate that smaller and low-lying glaciers in the tropical Andes are expected to disappear within a few decades under projected climate scenarios (Vuille et al., 2008). In recent studies about the future of Cordillera Vilcanota shows a scenario of the landscape mostly glacier-free until 2100 with an increment in glacial lake volume in the nearfuture (Drenkhan et al., 2018), as for Cordillera Huaytapallana even in the most optimistic scenario it is more likely glaciers would disappear by 2050 (López-Moreno et al., 2014), the Cordillera Ampato shows a decrease of 26.9% of glacier areas being large-scale circulation as ENSO and PDO phases what explain its maximum fluctuations which happens to others tropical glaciers in the Andes as well (Rabatel et al., 2013;Kozhikkodan-Veettil et al., 2016;Lamantia et al., 2024).Glaciers retreat has demonstrated affecting not only the dynamic of water supply and water demand but in socio-economic, environmental and cultural system discussed for several researchers (Salzmann et al., 2014;Zemp et al., 2015;Nussbaumer et al., 2017;Vuille et al., 2018;Condom et al., 2020), therefore, they highlight the importance of taking effective adaptive measures taking into account that one of the challenges faced by decision makers and researchers are the scarcity and availability of climatic, hydrological, in-situ meteorological and other necessary data as well as the creation of data sharing systems and platforms to be supported for the development of valid and quality information.To address this issue, in this report we present a database of four stations located in three Peruvian mountain ranges: Huaytapallana, located in the cordillera of the same name Huaytapallana; Coropuna, in the cordillera Ampato; Quelccaya and Quisoqupina, in the same cordillera called Vilcanota. These four stations have been recording hourly meteorological data for approximately 9 to 14 years including variables such as air temperature, wind speed and direction, precipitation, humidity and radiation, as detailed in the data description section. The data information presented has been processed and allowed to calculate daily, monthly, and annual averages / accumulation. This dataset has not yet been widely disseminated within the scientific community or the general public, and we hope it serves as a valuable resource for future research on tropical glaciers.Value of the dataAs it was stated, this dataset is being shared in order to deal with the scarcity of data in the tropics and high elevation locations such as these 4 glaciers. The value of these data of more than 10 years lies in the great utility they have for validation of satellite products and climate models, support for bias corrections, hydrological studies, glaciological studies and specific topics such as energy balance and mass balance in tropical glaciers since there are stations in ablation and accumulation zones of the glacier.3 MethodsIn 2011, the National Service of Meteorology and Hydrology of Peru (SENAMHI) through the Adaptation to the Impact of Rapid Glacier Retreat in the Tropical Andes Project (PRAA) financed by the World Bank's Global Environment Facility Fund installed two weather stations to monitor meteorological variables related to glacier mass loss and to study climate variations there. The first one is in Junin region on the front moraine of Lazontay glacier in the cordillera Huaytapallana at 4700 masl and the second one is in Cusco region, on the cordillera Vilcanota in the tongue of the Quisoquipina glacier at 5180 masl specifically in the ablation zone.In 2014, SENAMHI in cooperation with the Appalachia State University, a third station was also installed in Cusco region in the cordillera Vilcanota at 5560 masl on the accumulation zone of the Quelccaya ice field so in this case the purpose of this station is to monitor the meteorological variables involved in the accumulation processes, this station is also part of the Global Cryosphere Watch network since 2016. A year later, in 2015 a fourth station was installed in the cordillera Ampato in Arequipa region at an altitude of 5800 masl on the ablation zone of the Coropuna glacier plate tongue, which is the highest glacier volcano in the world.The Figure 1 shows the location of each station on the glaciers images of Landsat-8 as well as the in-situ automatic weather station equipment. The Quisoquipina station has a special design since it is located on the ablation zone of the glacier, in addition, considering that the glacier is constantly moving and that it cannot be visited more than once or twice a year, this station has been designed with a sensor stabilization system that allows maintaining an adequate perpendicular orientation of the radiation sensors to avoid obtaining erroneous data. The Coropuna station has a similar installation to the Quisoquipina station since it is also located in the ablation zone, so it has a stabilization system, both stations have been installed because of the need to study the meteorological variables that are involved in the ablation process, this station was installed in cooperation with the Specialized Association for Development (AEDES) responsible for the operation and maintenance of this. About the Quelccaya station, which is located in the accumulation zone near the mountain peak (5610 masl), was installed to study the meteorological variables related to the process of snow accumulation on the glacier, this station is static and maintained once a year. It is worth mentioning that this station was transferred from the University of Appalachia to SENAMHI in 2022. These three stations are located in places that are difficult to access, so it has been necessary to travel in vehicles for more than 4 hours, walk for 2 to 3 hours, and in some cases use horses to reach them. On the other hand, the Huaytapallana station was installed outside the glacier and it only presents security risks and thus is protected by metal fences; however, the other three stations, being located on the glacier, required special conditioning considering: glacier movement, frequency of visits to the station, security (access and theft). The four stations record data of recommended variables for CryoNet stations specifically for the surface meteorology (Global Cryosphere Watch, 2024): Air temperature, air humidity, wind speed and direction, air pressure, precipitation as well as short-wave and long-wave radiation and albedo; Coropuna station also counts with an ice surface temperature sensor. The recording time step is hourly and this information is being transmitted via satellite to SENAMHI headquarters.About the variables for glacier monitoring purposes, air temperature serves as an indicator of melting condition if values are greater than or equal to zero (≥0°C) and also could be used to compute the sensible heat flux and air humidity to compute latent heat flux. Wind speed is usually used to compute the sensible and latent heat flux meanwhile wind direction demonstrates occurrence of katabatic winds. The albedo of a surface also determines the amount of heat reflected by that surface. It is an essential parameter for understanding the thermal balance of a surface exposed to solar radiation, in the field of thermal comfort, climatology and spacecraft. 1 shows a list of equipment models for each station per variable, meanwhile details as its range and accuracy is attached as supplementary information. The Coropuna and Quisoquipina stations have a special structure consisting of a special tripod that allows keeping the radiation sensors in a horizontal position (short-wave and long-wave radiation incident perpendicularly) while moving due to the effect of mass loss in the ablation zone. At both stations the radiation measurement equipment is a Kipp & Zonen CNR-1 Net Radiometer (later changed to CNR-4 in 2017-2018) consisting of two pyranometers and two pyrgeometers arranged in pairs of each one facing upward and the others facing downward to monitor incoming and outgoing short-wave (albedo) and long-wave (heat flux) radiation. It was changed to CNR-4 net radiometer, since it is an improved version with a well-performing sensor as Dou et al. (2015) has demonstrated. On the other hand, it is worth to mention the sensors have not been observed to be covered by snow during maintenance visits to the stations.The raw dataset was collected as an excel file then it was cleaned using Python libraries. First, values as -999.9, NA or non-numeric were deleted, then a second filter were applied eliminating the values outside the plausible value limits as outlined in SENAMHI's (2022) quality control guidelines:Air Temperature:-40°C ≤ T ≤ 60°C Relative Humidity: 0.0 ≤ RH ≤ 100% Precipitation:0 ≤ PP ≤ 401mm/h Wind Speed: 0 ≤ WS ≤ 75m/s Wind Direction: 0° ≤ WD ≤ 360° Atmospheric Pressure: 300hPa ≤ P ≤ 1100hPa Incoming Shortwave Radiation: -1 ≤ SRin ≤ 1400W/m 2 Albedo: 0 ≤ Alb ≤ 1We also evaluate the time consistency and internal consistency of the parameters. Then, after a diurnal and seasonal boxplot analysis, it was applied an outlier detection to evaluate the "outer fences", defined by Turkey (1977) as Q1 (lower quartile) -3*IQR (inter-quartile range) and Q3 (upper quartile) + 3*IQR, and decided whether we discard them or not. Radiation variables were conditioned to fulfill the specific conditions according to its nature. In one hand, short-wave radiation components (outgoing and incoming) and albedo have non-null 12 values per day from 11:00 to 23:00 UTC which is the average sunlight period in local time (6 AM to 6 PM) and therefore the other values out of the range were replaced as zero, furthermore it was conditioned that outgoing values must be lower than incoming values, however, there are a lot of hours (6% of total data) where that does not happen in the Coropuna station, especially at hours near sunset and may be related to the residual radiation trapped in snow penitents which is characteristic of glaciers in arid areas as described by Lliboutry (1954) and Corripio and Purves (2005). On the other hand, for long-wave radiation, outgoing component can be greater or less than incoming values and about limits outgoing values can reach to 316 W/m 2 when temperature values are ≤ 0°C, values out of range were replaced by its respective limit. We also evaluated the radiation values following other data processing procedures of stations in snow as Nishimura (2023) and Wang (2018). Lastly, it was calculated Short-wave Net Radiation, Long-wave Net Radiation and Total Net Radiation as the sum of its components. This dataset is a compilation of hourly measurements from the four weather stations presented as a daily mean or cumulative sum (in the case of precipitation variable) in Figure 2, the start time of data recording is different for each station according to its date of installation.The Prospects for the future of the weather stations SENAMHI is currently planning to operate the stations within a Cryosphere observation network, which will be proposed to the Cryonet network operated by the WMO. In addition, it is planned to standardize the data acquisition, since the current frames are different.