Date of Award

May 2019

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Earth Sciences


Laura K. Lautz


Cordillera Blanca, Energy Balance Modeling, Groundwater, Peru, Stream temperature, Time-Lapse Infrared

Subject Categories

Physical Sciences and Mathematics


As climate change occurs, the availability and stability of water resources will be a global concern. The availability of reliable water resources will be of particular concern to regions that currently depend on meltwater from snowpack or glaciers during dry periods. One place whose water stability could be heavily impacted by the loss of meltwater resources is Peru. During the dry, arid months of May through September, the region to the west of the Peruvian Andes relies on stream water resources that originate in proglacial alpine catchments. The Cordillera Blanca, which is a mountain range in the Peruvian Andes, contains the highest density of tropical alpine glaciers worldwide, and the meltwater from these glaciers helps to sustain streamflow during the dry season. Unfortunately, tropical alpine glaciers are rapidly retreating as average annual air temperatures increase. In the Cordillera Blanca, glacial extent has decreased by more than 30% since 1930, and many of the glaciers have already passed a stage known as ‘peak water’, after which they continually contribute less and less water to streamflow. Recent research has indicated that although meltwater is a dominant contributor to streamflow during the dry season, groundwater within alpine aquifer systems may also be an important source of streamflow. Therefore, this research sought to investigate surface water – groundwater interactions in an alpine catchment of the Cordillera Blanca to gain a better understanding of the importance of groundwater in such regions.

This research focuses on the Quilcayhuanca Valley, which is a representative proglacial alpine catchment in the Cordillera Blanca, Peru. Glacial meltwater and groundwater contribute to streamflow within this catchment during the dry season. The Quilcayhuanca stream, along with streams that drain from adjacent catchments, flows into the Rio Santa which is a major stream in the region from which people withdraw water for a variety of uses. The high altitude wetlands in the Quilcayhuanca Valley and similar catchments, known as pampas, may represent an important source of groundwater to streamflow. The valley aquifers consist of a mixture of landslide and talus slope deposits from the steep, adjacent bedrock cliffs and glaciofluvial deposits. The valley aquifers are confined from above by glaciolacustrine sediments that were deposited when proglacial lakes were present. In order to better understand the surface water – groundwater interactions in such a setting, we have combined energy balance modeling of stream heat fluxes, thermal infrared remote sensing of stream temperatures, and groundwater flow modeling to a section of the Quilcayhuanca stream that is downstream of direct glacial melt inputs to examine the influence of groundwater.

Energy balance modeling of stream temperatures, also known as heat tracing, can be used to estimate groundwater contribution to a stream. This method uses the fluxes of energy into and out of a stream to calculate stream temperature through time and space, and the difference between calculated and observed stream temperatures can be attributed to groundwater inflow. Meteorological and longitudinal stream temperature data were recorded for approximately a week in a portion of the Quilcay stream, and used as input for an energy balance model of the reach. Various model inputs were also varied in order to assess the sensitivity of the model to certain parameters and determine the extent to which uncertainty in certain model parameters affects estimates of groundwater influx. An input that was investigated in depth was the extent to which uncertainty in the daily diurnal streamflow signal affects groundwater inflow estimates, since the streamflow in proglacial streams varies diurnally due to glacial melt. Groundwater influx to the model reach was estimated at 42.1 L s−1 km−1, and the uncertainty in diurnally fluctuating streamflow was determined to affect the estimated relative groundwater contribution to streamflow by approximately ±5%.

In order to improve the spatial resolution of the stream temperature data used to inform the energy balance modeling of stream temperatures, we explored the use of ground-based, time-lapse infrared remote sensing to measure stream temperatures along the same study reach in the Quilcayhuanca Valley. During two field seasons, a thermal infrared (TIR) camera was deployed on the steep valley cliffs, recording time-lapse images of stream temperature. Analysis of the infrared images revealed that measured infrared stream temperatures are highly sensitive to infrared temperatures from other objects in the environment that reflect off the stream surface, often leading to large discrepancies between in-situ and remote temperature measurements. We determined that even at nadir views, reflected temperatures can still affect measured TIR stream temperatures, and that previous analytical correction methods performed by the hydrology community have not accurately represented reflected infrared temperatures. While such analytical correction methods could be improved through more accurate measurements of reflected temperatures, an empirical correction approach can also be used to correct stream infrared temperatures. While this data was not ultimately used to refine the stream temperature energy balance model of the Quilcay stream, this investigation has helped improve our understanding of remote sensing of stream temperatures using ground-based, time-lapse thermal infrared imagery.

To complement the groundwater influx rates to the Quilcayhuanca stream estimated by stream temperature energy balance modeling, a groundwater flow model of the same pampa aquifer system was developed. Precipitation, stream stage, and groundwater level data that had been collected over various time periods were compiled to parameterize and calibrate the model. Numerous model simulations were explored to determine which model configuration best reproduced the annual hydraulic head patterns in the aquifer and the estimated dry season groundwater flux to the reach of the Quilcay stream. The modeled groundwater flux estimates were then used to estimate the amount of groundwater entering the stream from all pampa regions of the catchment above the gauging station. Results indicate that about 7-53% of Quilcayhuanca streamflow is derived from groundwater depending on the month during the dry season. As the dry season progresses, the relative contribution of groundwater to the stream decreases as the aquifer becomes depleted. Travel time analysis also indicates that the residence time of water in the pampa aquifer system is relatively short, with >80% of water moving through the system in 6 months and the remaining water exiting after around 1-1.5 years. These results suggest that groundwater within these proglacial alpine catchments also contributes to streamflow, and that streamflow is vulnerable glacial loss, especially at the end of the dry season.


Open Access