1. Background
Mountain headwater regions provide much of the world’s water (Chang,
Foster, Hall, Rango, & Hartline, 1982; Ikeda et al., 2010) and in the
western United States, the Colorado River provides water for 30 million
people and 4 million acres of irrigated agriculture (Reclamation, 2012).
However, many western watersheds lose more water to the atmosphere
through evaporation and transpiration (ET) than is discharged into
rivers and streams (Knight, Fahey, & Running, 1981; Trenberth, Smith,
Qian, Dai, & Fasullo, 2007). In the Western U.S., ET is often the
second largest component of the water balance, after precipitation, and
is also associated with a watershed’s energy balance making accurate
estimations of this flux vitally important to overall water availability
predictions (Healy, Winter, LaBaugh, & Franke, 2007) as well as
regional energy budgets. A recent study by Milly and Dunne (2020)
estimates that the mean annual discharge of the Colorado River is
decreasing due to increased evapotranspiration while another study shows
similar trends in the Alps with an air temperature increase causing an
increase in ET leading to a decrease in runoff consistent with a 3%
precipitation decrease (Mastrotheodoros et al., 2020). Both studies
reveal the importance of ET flux to overall water availability. However,
ET is challenging to quantify in headwaters regions as these regions are
often marked by complexities such as heterogeneous land cover, large
topographic gradients, and spatially varying atmospheric patterns. Due
to these complexities, few studies have tested the effectiveness of
measuring ET in complex headwaters regions (Flerchinger, Marks, Reba,
Yu, & Seyfried, 2010).
Eddy covariance is a commonly used technique for quantifying latent
heat—equivalent to ET multiplied by the latent heat of vaporization of
water at a given temperature— and often used interchangeably with ET
in this paper (Ceperley et al., 2017; Hirschi, Michel, Lehner, &
Seneviratne 2017; Maes, Gentine, Verhoest, & Miralles, 2019; Mamadou et
al., 2014). This method computes the flux as the covariance between
deviation from the mean of the vertical wind speed and deviation from
the mean of the concentration multiplied by the air density (Burba,
2013). This method is effective in flat terrain with homogenous land
cover due to the assumptions that must be made using this technique,
including the lack of surface heterogeneity, adequate fetch, and
turbulent fluxes (Burba & Anderson, 2006). This then leads to
difficulty in predicting how these water and energy fluxes, and overall
water availability, will change with a changing climate.
Despite the challenges outlined above, some studies have successfully
used the eddy covariance in high-elevation environments. Burns, Blanken,
Turnipseed, Hu, & Monson (2015) studied how changes in precipitation
affect the overall surface energy budget in a high-elevation forest
while Frank, Massman, Ewers, Huckaby, & Negrón (2014) studied how
disturbances, such as bark beetles, shift the energy balance components
in the Rocky Mountains. Complimentary studies have also used eddy
covariance in western riparian areas leading to helpful end-member ET
values, which range from 10 to 30% surface energy balance closure error
(Nagler et al., 2005; Scott et al., 2004; Scott et al., 2008; Wilson et
al., 2002). Through the measurement of these fluxes and thus better
estimations of the water and energy balances, these studies allow for a
better understanding of energy fluxes in mountain environments and
result in data sets that can be used to check model performance.
In this study, we improve the quantification of water and energy fluxes
in high-elevation, complex systems, we estimate groundwater
contributions to ET, and we compare our results to other sites in the
Rocky Mountain region. Our study site, a basin that contributes to the
Upper Colorado River Basin and eventually the Colorado River, provides
an archetype of other Rocky Mountain basins. Our study period includes
multiple water years characterized by both wet and dry conditions
resulting in ET estimations useful for application across the entire
basin and to other headwaters regions. This study leads to greater
understanding of water availability in headwaters systems, which can
then be used for ecological and hydrological applications and downstream
water planning.