Doctor of Philosophy (PhD)
First Committee Member
Second Committee Member
Third Committee Member
Fourth Committee Member
Fifth Committee Member
Number of Pages
Desert soils cover about one third of the Earth’s land surface. Despite their large extent, however, our understanding of desert soil hydraulic processes and properties is still rather limited. In particularly with respect to the near-surface (top centimeters) of the soil profile, which play a critical role for infiltration, redistribution and evapotranspiration of rainwater, the limiting resource for life in desert environments, but also hosts most of the biologic activity and controls runoff, erosion as well as the emission of dust (Nannipieri et al.,2003; Bradford et al.,1987). Additionally, deserts are ideal locations for electricity generation with solar energy using large-scale photovoltaic (PV) facilities with considerable impacts on desert environments. To minimize these impacts (Sinha et al. 2018) a better understanding is needed how facility-scale solar PV installations may affect the local hydrology, in particular the moisture distribution in the soil underneath and between rows of solar panels.
A recent study by Dijkema et al. (2018) introduced a modeling framework to simulate the moisture dynamics of bare, desert soil. Their model was able to capture water redistribution under infiltration as well as evaporation conditions. For the latter, however, only when the soil was relatively moist (volumetric moisture content >10%, matric head 2). For dryer soil conditions, however, as they occur between rainfall events and are typical for desert soils most of the time, the model consistently underestimated evaporative fluxes and, subsequently, overestimated soil moisture content. The primary goal of this study therefore was to improve the model by Dijkema et al. (2018) by using the Peters-Durner-Iden (or PDI) water retention and hydraulic conductivity functions (Peters, 2013, 2014), which can capture not only capillary but also film flow of liquid water in the soil pores, to better simulate the moisture dynamics of bare, near-surface desert soils.
I found that the PDI water retention functions better represent measured water retention curves than the bimodal van Genuchten (or BVG) water retention functions used by Dijkema et al. (2018). In particular for the critical range of volumetric moisture contents between 5% and 10% (corresponding saturation degrees between 16% and 32%, and matric heads between pF 2 and pF 4). The BVG water retention functions predicted higher suctions (lower matric heads) than the PDI functions for volumetric moisture contents in the range from 5% to 10%, likely underestimating the hydraulic conductivity values and water fluxes in the soil for volumetric moisture contents ranging from 5% to 10%. Interestingly, the BVG and PDI hydraulic conductivity functions were not noticeably different within the range between pF 2 and 3. For pF values higher than 3, however, the PDI hydraulic conductivity functions predicted higher hydraulic conductivity values than the BVG hydraulic conductivity functions. Therefore, the hypothesis put forward by Dijkema et al. (2018) that including film flow may improve model predictions could be confirmed for pF >3. For pF values between 2 and 3, however, is probably rather the difference between the BVG and PDI water retention functions than the difference between the BVG and PDI hydraulic conductivity functions that led to the improved soil moisture simulations. Using PDI instead of BVG hydraulic functions, the improved model by Dijkema et al. (2018) (hereafter referred to as Luo et al. 2019a) was able to much better simulate the measured soil moisture data from the SEPHAS Lysimeter 1 (https://www.dri.edu/sephas) at 10 cm depth and below.
To further test the model by Luo et al. (2019a), I compared simulated with measured moisture content data from the top 50 mm of SEPHAS Lysimeter 1 soil profile that was instrumented with an array of Triple-Point-Heat-dissipation Probes (TPHP, East 30 Sensors, Inc., Pullman, WA). The TPHPs allow to monitor volumetric moisture content and temperature near the soil surface from 6 mm to 54 mm depth at 6 mm vertical intervals on an hourly basis. The TPHPs provided a unique dataset that captured the moisture distribution of the near surface soil for nearly a decade. The measured soil moisture dataset showed that only the top 20 mm of the soil dry out completely (i.e. reach a moisture content below the detection limit of the TPHP) when no precipitation occurs for several months such as during summer and fall 2019. The soil moisture contents between 20 to 50 mm depths, however, converge towards some equilibrium value ranging between 3.5% and 5% (or 15% to 21% saturation, respectively). I compared measured soil moisture time series for different depth with simulations using the model by Luo et al. (2019a) and found that the latter is able to forward simulate quite accurately the soil moisture redistribution in the top 50 mm of the soil profile for a wide range of soil moisture contents ranging from 15% to 95% saturation using only independently determined soil physical parameters as well as precipitation and evaporation as flux boundary conditions. Measurements and simulations show that moisture from the top 50 mm of the soil profile evaporates back into the atmosphere within one month maximum after a rain event. The spatial and temporal moisture distributions in the top 50 mm of the soil profile during evaporation point toward a two-step evaporation process as proposed by Shokri et al. (2009) and Or et al. (2013), which could not be captured by Luo et al. (2019a).
In a third step, I applied the model by Luo et al. (2019a) to explore the impact of solar arrays on the moisture distribution in the soil of a PV solar facility with its characteristic rows of solar panels that change the way rainwater reaches the soil and, therefore, likely also the local hydrology at the solar facility scale. I was particularly interested in how solar arrays may lead to concentrated infiltration of rainwater into the soil and whether this change of infiltration pattern has an impact on the water balance of the soil within the solar facility. For this purpose, I set up a process-based soil physics model with HYDRUS-2D for an experimental site within of the “Solar 1” PV solar plant at the Desert Research Institute (DRI) in Las Vegas (http://www.dri.edu/renewable-energy#Solar-Generation). Solar 1 consists of 204 PV solar panels, has a nominal output of 54 kW, and produces electricity for DRI’s Las Vegas campus. Soil physical processes were simulated based on Luo et al. (2019a) and soil physical properties were determined from soil samples collected at Solar 1. Model calculations were driven by measured precipitation and calculated evaporation data. Evaporation was calculated from measured air-temperature, net radiation, relative humidity and wind speed data from the nearby CEMP station at DRI Las Vegas (https://cemp.dri.edu/cgibin/cemp_stations.pl?stn=lasv). Measured water content values from 5 cm, 15 cm and 25 cm depth in the drip line, between panels and underneath panels were then compared with HYDRUS-2D simulations.
The HYDRUS-2D simulations showed that the solar PV panels concentrate rainfall along the drip lines of the panels, which causes deeper infiltration of rainwater along the drip lines compared to areas between rows and no infiltration underneath a row of solar PV panels. This finding is in accordance with recent lysimeter studies on infiltration into and evaporation off bare, arid soil (Koonce, 2016; Lehmann et al., 2019) that shown that the deeper rainwater infiltrates into the soil, the less likely it is to evaporate back into the atmosphere. HYDRUS-2D calculations agreed best with measured moisture content values from the soil at the drip line and underneath the panels. The former is quite remarkable since the drip line is the area with the highest moisture dynamics. The latter is less surprising since not much soil moisture change occurs underneath the panel. The HYDRUS-2D model, however, had only limited success in simulating the moisture content of the soil between the panels. A possible reason is that the soil between the panels is more heterogenous than assumed by the HYDRUS-2D model. It was also noticed that during rain events water tend to pond on the soil between panels, which HYDRUS-2D cannot take into account but may also considerably affect infiltration and water redistribution of rainwater within the soil between panel.
In conclusions, the model by Luo et al. (2019a) can accurately simulate moisture content values as low as 3.5% (corresponding to 15% saturation and pF 4.7) when using the PDI instead of the BVG water retention and hydraulic conductivity functions. Provided good quality water retention and hydraulic conductivity data are available to parametrize the PDI functions. I.e. the model by Luo et al. (2019a) is able to simulate the moisture dynamics of a bare, near-surface desert soil from near-saturation to about twice the air-dry moisture content. The study also shows how the soil moisture distribution in the top 50 mm of a bare desert soil changes as a function of individual precipitation events as well as during several month without rainfall. HYDRUS-2D simulations in combination with measurements showed how rows of PV solar panels affect the soil moisture distribution between and underneath the panels and how initial, or antecedent, moisture content plays a role in terms of infiltration pattern.
Further research is needed to explore whether the PDI model could capture water redistribution at volumetric soil moisture contents even lower than 3.5% (4.7), maybe as low as 2%, which would correspond to air-dry conditions for Lysimeter 1 soil assuming an average air relative humidity of 60% (at 25˚C). The PDI model (Peters, 2013) also includes a vapor flow component, which has not been taken into account for this study but could be employed to simulate water flow under even dryer conditions t. In terms of water redistribution due to PV solar panels, a follow up study that simulates a series of storms events could shed light on whether areas of concentrated infiltration in the drip line indeed develop into conduits of deeper infiltration, especially for smaller storm events, and therefore might have a profound impact on the water balance of a soil under arrays of solar panels. Moisture measurements in the soil of the dripline as well as between the rows of solar panels shows a small but potentially important difference in moisture content, especially considering how sensitive soil hydraulic conductivity is on soil moisture content. Overall, this study has improved our understanding and ability to simulate the moisture dynamic of bare, near-surface desert soils and may help to guide human activities in the desert such as renewable energy generation in a more sustainable direction by minimizing their impact on fragile desert environments.
University of Nevada, Las Vegas
Luo, Yuan, "Moisture Dynamics of a Near-Surface Desert Soil" (2019). UNLV Theses, Dissertations, Professional Papers, and Capstones. 3822.
IN COPYRIGHT. For more information about this rights statement, please visit http://rightsstatements.org/vocab/InC/1.0/