Monitoring precipitation in mountainous areas using traditional tipping-bucket rain gauges (TPB) has become challenging in sites with strong variations of air temperature and wind speed (Ws). The drop size distributions (DSD), amount, and precipitation-type of a Parsivel OTT2 disdrometer installed at 4730 m above sea level (close to the 0 °C isotherm) in the glacier foreland of the Antisana volcano in Ecuador are used to analyze the precipitation type. To correct the DSDs, we removed spurious particles and shifted fall velocities such that the mean value matches with the fall velocity–diameter relationship of rain, snow, graupel, and hail. Solid (SP) and liquid precipitation (LP) were identified through −1 and 3 °C thresholds and then grouped into low, medium, and high Ws categories by k-means approach. Changes in DSDs were tracked using concentration spectra and particle’s contribution by diameter and fall velocity. Thus, variations of concentration/dispersion and removed hydrometeors were linked with Ws changes. Corrected precipitation, assuming constant density (1 g cm−3), gives reliable results for LP with respect to measurements at TPB and overestimates SP measured in disdrometer. Therefore, corrected precipitation varying density models achieved fewer differences. These results are the first insight toward the understating of precipitation microphysics in a high-altitude site of the tropical Andes.

Tropical glaciers are excellent indicators of climate variability due to their fast response to temperature and precipitation variations. At same time, they supply freshwater to downstream populations. In this study, a hydro-glaciological model was adapted to analyze the influence of meteorological forcing on melting and discharge variations at Glacier 12 of Antisana volcano (4,735–5,720 m above sea level (a.s.l.), 1.68 km2, 0°29′S; 78°9′W). Energy fluxes and melting were calculated using a distributed surface energy balance model using 20 altitude bands from glacier snout to the summit at 30-min resolution for 684 days between 2011 and 2013. The discharge was computed using linear reservoirs for snow, firn, ice, and moraine zones. Meteorological variables were recorded at 4,750 m.a.s.l. in the ablation area and distributed through the altitudinal range using geometrical corrections, and measured lapse rate. The annual specific mass balance (−0.61 m of water equivalent -m w.e. y−1-) and the ablation gradient (22.76 kg m−2 m−1) agree with the values estimated from direct measurements. Sequential validations allowed the simulated discharge to reproduce hourly and daily discharge variability at the outlet of the catchment. The latter confirmed discharge simulated (0.187 m3 s−1) overestimates the streamflow measured. Hence it did not reflect the net meltwater production due to possible losses through the complex geology of the site. The lack of seasonality in cloud cover and incident short-wave radiation force the reflected short-wave radiation via albedo to drive melting energy from January to June and October to December. Whereas the wind speed was the most influencing variable during the July-September season. Results provide new insights on the behaviour of glaciers in the inner tropics since cloudiness and precipitation occur throughout the year yielding a constant short-wave attenuation and continuous variation of snow layer thickness.

The precipitation phase (PP) affects the hydrologic cycle which in turn affects the climate system. A lower ratio of snow to rain due to climate change affects timing and duration of the stream flow. Thus, more knowledge about the PP occurrence and drivers is necessary and especially important in cities dependent on water coming from glaciers, such as Quito, the capital of Ecuador (2.5 million inhabitants), depending in part on the Antisana glacier. The logistic models (LM) of PP rely only on air temperature and relative humidity to predict PP. However, the processes related to PP are far more complex. The aims of this study were threefold: (i) to compare the performance of random forest (RF) and artificial neural networks (ANN) to derive PP in relation to LM; (ii) to identify the main drivers of PP occurrence using RF; and (iii) to develop LM using meteorological drivers derived from RF. The results show that RF and ANN outperformed LM in predicting PP in 8 out of 10 metrics. RF indicated that temperature, dew point temperature, and specific humidity are more important than wind or radiation for PP occurrence. With these predictors, parsimonious and efficient models were developed showing that data mining may help in understanding complex processes and complements expert knowledge.