Assessing Riparian Evapotranspiration Dynamics in a Water Conflict Region in Nebraska, USA

Abstract

The escalating pressure on water resources in agricultural regions has become a catalyst for water conflicts. The adoption of innovative approaches to estimate actual evapotranspiration (ETa) offers potential solutions to mitigate conflicts related to water usage. This research presents the application of a remote sensing-based methodology for estimating actual evapotranspiration (ETa) based on a two-source energy balance model (TSEB) for riparian vegetation in Nebraska, US using the Spatial EvapoTranspiration Modeling Interface (SETMI). Estimated results through SETMI and field data using the eddy covariance system (EC) considering the period 2008–2013 were used to validate the energy balance components and ETa. Modeled energy balance components showed a strong correlation to the ground data from EC, with ET presenting R2 equal to 0.96 and RMSE of 0.73 mm.d⁻¹. In 2012, the lowest adjusted crop coefficient (Kcadj) values were observed across all land covers, with a mean value of 0.49. The years 2013 and 2012, due to the dry conditions, recorded the highest accumulated ETa values (706 mm and 664 mm, respectively). Soybeans and corn exhibited the highest ETa values, recording 699 mm and 773 mm, respectively. Corn and soybeans, together accounting for a substantial portion of the land cover at 15% and 3%, respectively, play a significant role. Given that most fields cultivating these crops are irrigated, both pumped groundwater and surface water directly impact the water source of the Republican River. The SETMI model has generated appropriate estimated daily ETa values, thereby affirming the model’s utility as a tool for assisting water management and decision-makers in riparian zones.

1. Introduction

Water conflicts and population growth are interconnected challenges that pose significant threats to global sustainability, particularly in regions where agriculture heavily depends on irrigation. Sustainable irrigation practices and efficient water management are crucial to mitigating conflicts for ensuring water and food security. Balancing the needs of various sectors and implementing efficient water management strategies are key components for addressing these challenges.

The water conflict in the Republican River Basin involves Nebraska, Kansas, and Colorado and centers around the allocation and usage of water from the Republican River. The dispute primarily revolves around adherence to the Republican River Compact, an agreement reached in 1943 to ensure equitable sharing of water resources among the three states [1]. The conflict arises when one state extracts more water from the Republican River than allowed under the compact, impacting downstream states, particularly during times of drought or when water resources are limited. The disputes have involved complex hydrological and legal considerations, including debates over measurement methods, water accounting, and compliance with the compact. Efforts have been made to find collaborative solutions and improve water management practices to prevent future conflicts.

Estimating evapotranspiration (ET) is crucial for effective irrigation management as it provides valuable information about gross crop irrigation requirements helping farmers and irrigation managers optimize water use by ensuring that the right amount of water is applied to meet the specific needs of crops. Over-irrigation can lead to wastage and environmental issues, while under-irrigation can result in reduced crop yields. In recent years, remote sensing using satellite sensors has become an increasingly important tool to estimate ET and has several advantages over traditional ground-based methods providing spatial information over large areas. Further, allowing daily ET monitoring at a low cost [2,3,4].

Riparian areas are hydrologically connected zones adjacent to streams, rivers, or wetlands, where interactions between surface and groundwater control soil moisture, vegetation, and evapotranspiration. They act as key interfaces regulating water and nutrient exchanges between terrestrial and aquatic systems [5,6]. In managed landscapes, riparian vegetation may also include agricultural crops or pastures influenced by shallow groundwater or flooding, which still perform essential hydrological functions [7]. Accurate assessment of ET in riparian zones is critical for understanding hydrological processes that govern water balance. Riparian zones function as key transition areas where vegetation-mediated ET influences groundwater recharge, streamflow regulation, and the overall connectivity between surface and subsurface hydrological components. Quantifying ET in these environments allows for the evaluation of water use efficiency, ecosystem resilience, and feedback between land management practices and riverine hydrodynamics, thereby supporting more sustainable water resource and riparian management strategies. Because riparian zones often extend across large and heterogeneous landscapes, remote sensing approaches provide an effective means to quantify spatially distributed ET patterns that cannot be captured through ground-based methods alone. Such assessments are critical for evaluating the impacts of land use on riparian water resources and ecosystem sustainability.

Among several remote sensing-based ET models, we can cite the hybrid model Spatial EvapoTranspiration Modeling Interface (SETMI) developed by [8,9]. The SETMI approach can estimate ET in three different ways: (a) using the two-source energy balance model (TSEB) based on multispectral and thermal images and instantaneous weather datasets to estimate actual ET. Using the Two-Source Energy Balance (TSEB) model allows separate estimation of soil and canopy fluxes, improving ET accuracy under partial canopy cover and heterogeneous conditions common in rainfed fields [10]; (b) using the basal crop coefficient as a function of vegetation indices derived from the canopy spectral reflectance values along with reference evapotranspiration from weather station data to determine the actual crop ET which is then used to estimate the remote sensing-based soil water balance (RSWB); (c) the hybrid approach based on coupling the TSEB and the basal crop coefficient approaches and assimilating the actual ET from the TSEB into the RSWB and update the soil water content if needed. The SETMI model interpolates and extrapolates ET between satellite acquisition dates, improving the maintenance of a soil water balance in the crop root zone.

With growing concerns about water scarcity and the need for sustainable agriculture, efficient water management has become paramount. Accurate ET estimates enable farmers to use water resources more judiciously, conserving this precious resource for both current and future generations. In this study we aimed to estimate ET using the SETMI model for the riparian vegetation and surrounding areas along the Republican River in the state of Nebraska, covering the period from 2008 to 2013.

2. Material and Methods

2.1. Study Site

Evapotranspiration (ET) was modeled for an approximately 350 km long swath of riparian land along the Republican River in southwestern Nebraska during 2008–2013, coinciding with the availability of eddy covariance (EC) flux data used to validate the remote sensing model (Figure 1).

Nebraska is in the midwestern USA and lies in the Great Plains region over the Ogallala aquifer, one of the largest aquifers in world. Therefore, Nebraska is well known for the largest irrigated agriculture and livestock production in the nation. Corn is the most widely grown cropland and soybean is the second largest harvested crop in Nebraska and often is cultivated in rotation with corn. The annual accumulated precipitation in Nebraska, based on weather data collected from 1981 to 2010 (standard 30-year normal), varies across the state from western Nebraska (435 mm) to eastern Nebraska (760 mm) and is irregular over the year [11]. According to the Köppen classification [12], the study area presents a climate type defined as Dfa (humid continental climate), with cold winters and relatively dry and hot summers. Records from the weather station for the US National Weather Service located in the study site indicate that for the last 30 years (1981–2010), statistics show an average annual temperature equal to 11 °C, average maximum temperature equal to 19.1 °C, average minimum of 4.8 °C, and average precipitation of 632 mm year⁻¹ [11]. Precipitation is unevenly distributed over the year with most of the precipitation occurring during the spring and summer, and relatively drier conditions in fall and winter. The study site has high soil variability along the Republican River with deep silty clay loams, sandy and well drainage soils [13].

2.2. SETMI Model

The hybrid approach within the model was not used due to the impossibility of obtaining the required daily as applied irrigation and precipitation mainly for irrigated corn and soybeans fields, so only the TSEB in the SETMI model was used in this study. The TSEB model [10] as formulated in the SETMI model [8] with modifications is described by [14].

In the TSEB, the soil and plant contributions to energy fluxes are considered separately rather than as a combined surface (hence two sources), following the general Equation (1).

(R_nc + R_ns) = (H_c + H_s) + (LE_c + LE_s) + G

(1)

R_n is the net radiation (w m⁻²); H is the sensible heat flux (w m⁻²); LE is the latent heat flux (w m⁻²); G is the soil heat flux (w m⁻²). Subscripts C and S indicate canopy and soil, respectively.

The R_ns and R_nc are estimated following [15] and, G according to [16], (Equation (2)).

G = Ag × R_ns

(2)

R_ns is the soil net radiation (w m⁻²); G is the soil heat flux (w m⁻²). the rate of heat transfer between the soil surface and the deeper layers of the ground; Ag is the ratio of G to R_n [17].

An initial value of canopy sensible heat flux H_c is calculated using the Priestly Taylor (PT) approximation [10], Equation (3).

$${\mathrm{H}}_{\mathrm{C}}=\left[1-{\mathsf{\sigma }}_{\mathrm{P}\mathrm{T}}{\mathrm{f}}_{\mathrm{g}}{\displaystyle \frac{\Delta }{\Delta +\mathsf{\gamma }}}\right]{\mathrm{R}}_{\mathrm{n}\mathrm{c}}$$

(3)

H_c is the sensible heat flux from canopy (w m⁻²); R_nc is the net radiation from canopy (w m⁻²); Δ is the slope of the vapor pressure temperature curve (mb K⁻¹) [18]; σ_PT is the Priestly Taylor coefficient (1.26); f_g is the fraction of green; γ is the psychrometric constant (0.665 mb K⁻¹).

An iteration loop is started for α_PT [19], or R_c. The maximum number of iterations is set to 100. A loop to iterate for T_c since T_c is used to calculate since T_c is used to calculate Δ and γ* this follows [20]. The canopy latent heat flux (LE_c) is then calculated following Equation (4).

$${\mathrm{L}\mathrm{E}}_{\mathrm{c}}={\mathsf{\sigma }}_{\mathrm{P}\mathrm{T}}{\mathrm{f}}_{\mathrm{g}}{\displaystyle \frac{\Delta }{\Delta +\mathsf{\gamma }}}{\mathrm{R}}_{\mathrm{n}\mathrm{c}}$$

(4)

LE_c is the latent heat flux from canopy (w m⁻²); σ_PT is the Priestly Taylor coefficient; f_g is the fraction of green; Δ is the slope of the vapor pressure temperature curve (mb K⁻¹); γ is the psychrometric constant (mb K⁻¹); R_nc is the net radiation from canopy (w m⁻²).

H_c is then redundantly recalculated using a rearrangement of Equation (5) in [10].

H_c = R_nc − LE_c

(5)

H_c is the sensible heat flux from canopy (w m⁻²); R_nc is the net radiation from canopy (w m⁻²); LE_c is the net radiation from canopy (w m⁻²).

The soil component of sensible heat flux (H_s) is calculated following Equation (6).

$${\mathrm{H}}_{\mathrm{s}}={\displaystyle \frac{\left({\mathrm{T}}_{\mathrm{s}}-{\mathrm{T}}_{\mathrm{a}\mathrm{c}}\right){\mathrm{c}}_{\mathrm{p}\mathrm{m}}{\mathsf{\rho }}_{\mathrm{m}}}{{\mathrm{R}}_{\mathrm{s}}}}$$

(6)

H_s is the sensible heat flux from soil (w m⁻²); T_s is the soil temperature (K); T_ac is the air temperature within the canopy (K); C_pm is the specific heat of moist air (J kg⁻¹ K⁻¹), ρ_m is the density of the moist air (kg m⁻³); R_s is the soil resistance.

The latent heat flux from the soil is calculated following [10], Equation (7).

LE_s = R_ns − G − H_s

(7)

LE_s is the net radiation from soil (w m⁻²); R_ns is the net radiation from soil (w m⁻²); G is the soil heat flux (w m⁻²); H_s is the sensible heat flux from soil (w m⁻²).

Instantaneous latent heat flux (LE) computed using the TSEB was scaled to a daily actual ET value following [21] using the ratio of instantaneous and daily reference ET according to Equation (8).

$${\mathrm{E}\mathrm{T}}_{\mathrm{a}}={\mathrm{L}\mathrm{E}}_{\mathrm{i}}\left({\displaystyle \frac{3600}{\mathsf{\lambda }}}\right)\left({\displaystyle \frac{{\mathrm{E}\mathrm{T}}_{\mathrm{r},\mathrm{d}}}{{\mathrm{E}\mathrm{T}}_{\mathrm{r},\mathrm{i}}}}\right)$$

(8)

ET_a is the actual evapotranspiration (mm per day); ET_r is the reference evapotranspiration (mm; alfalfa reference); LE is the latent heat flux (w m⁻²); λ is the latent heat of vaporation (w m⁻²) [22]. Subscripts d and i indicate daily and instantaneous, respectively.

Five dominant land cover classifications based on [23] and available at https://s.cnmilf.com/user74170196/https/croplandcros.scinet.usda.gov/ (access date: 5 October 2025), comprising over 80% of riparian vegetation within the swath width, were considered, including cottonwood, eastern red cedar, maize, and short grass prairie (along with other grasses). Short grasses account for 50% of the riparian vegetation land use, followed by corn at 15%, cottonwood at 6%, bare soil at 4%, soybean at 3%, and red cedar at 2%. The remaining 20% is represented by roads, open water, and small crop fields. Landsat 5, 7, and 8 Level 1 thermal infrared imagery and Level 2 surface reflectance imagery were used as the primary remote sensing inputs. We used “pixel_qa band” to avoid cloud or haze conditions and guarantee pixel quality on the study site. Landsat imagery from March through November for 2008–2013 was considered for a Landsat scene that covered this entire reach of the river. In total, 50 Landsat scenes were used in the TSEB modeling.

The biophysical property relationships for the five land covers used within SETMI are shown in

Table 1. Notable two-source energy balance model parameters and sources used in SETMI.

Vegetation Type	LGA		LSA		SR		TE			LS 1	Hc		Hc:W 1
	VIS 1	NIR 1	VIS 1	NIR 1	VIS 2	NIR 2	Soil 1	Veg 1	Sen Veg 3		Min	Max
Cottonwood	0.86	0.37	0.84	0.61	0.15	0.25	0.96	0.98	0.95	0.1	24	24	1
Easter Red Cedar	0.89	0.60	0.84	0.61	0.15	0.25	0.96	0.98	0.95	0.05	8	8	1
Grasses	0.82	0.28	0.42	0.04	0.15	0.25	0.96	0.98	0.95	0.02	0.1	0.5	1
Maize	0.83	0.35	0.49	0.13	0.15	0.25	0.96	0.98	0.95	0.2	--- 4	--- 4	1
Soybean	0.85	0.20	0.49	0.13	0.15	0.25	0.96	0.98	0.95	0.2	--- 4	--- 4	1
Bare Soil	0.1	0.28	0.42	0.04	0.15	0.25	0.96	0.98	0.95	0.02	0.1	0.1	1

Note(s): ¹ [8]; ² [20]; ³ [24]; ⁴ [25], except during the offseason, was 0.1. LGA: leaf green absorptivity. LSA: leaf senescence absorptivity. SR: soil reflectance. TE: Thermal emissivity. LS: leaf size. Hc: canopy height (m). W: width. VIS: visible. NIR: near infrared.

including leaf green and senescence absorptivity on visible and near infrared spectrum, soil reflectance, thermal emissivity, leaf size, canopy height and the ratio canopy height and width.

The necessary weather data for model input were obtained from the Nebraska Mesonet McCook 4NE weather station near the study area. Instantaneous and daily weather data were obtained from the High Plains Regional Climate Center and the ASCE Standardized Penman–Montieth equation [26] was used to estimate reference evapotranspiration (ETr) for an alfalfa reference crop. The crop coefficient (Kc) was determined by calculating the ratio between the modeled daily ETa from SETMI (considering all the image output pixels) and the daily ETr according to [18]. Given that the soil water content during the study period likely did not have optimal conditions due to inadequate precipitation distribution and amount throughout the period, particularly in the rainfed crop fields (corn and soybean), we assumed an adjusted Kc (Kc_adj). Daily Kc_adj was estimated using daily SETMI-ETa24 h (satellite overpasses day), annually considering the entire period and each land cover, divided by ETr from the weather station following [18], then Kc_adj values from satellite overpasses day were interpolated using day of year to estimate daily Kc for the entire growing season. With the daily Kc_adj, it was possible to estimate daily ETa multiplying daily Kc_adj to ETr.

2.3. Surface Energy Balance and ET Validation

Validation energy flux data were obtained from an EC flux system on natural grass land (40°01′54″ N; 101°33′56″ W, 91 m above sea level) to measure the micrometeorological variables and the energy balance components for estimating evapotranspiration. The EC system consisted of a three-dimensional sonic anemometer and an infrared gas analyzer—IRGASON, positioned considering the prevailing wind direction, operated by data logger (CR 3000, Campbell Scientific Instruments, Logan, UT, USA) to record raw high-frequency data at 10 Hz measured at 3.0 m above the ground surface.

The micrometeorological variables measured above the canopy were net radiation (CNR4) and precipitation (CS700-L). This equipment was fixed at 3.0 m above the soil surface. On the ground, heat flux plates were installed to measure the heat flow in the soil (HFP01-L), the data were collected continuously at 5 s intervals averaging over 30 min.

The raw data from the EC system (10 Hz) were processed using EddyPro Advanced software (version 7.0.9) and Tovi software (version 2.9.1) was used for data gap filling and flow partitioning every 30 min.

The Bowen Ratio with the flux tower data was used to adjust λE and H by forcing the closure following the procedure suggested by [27], the flux data obtained showed an energy balance closure on a half hourly scale of about 87%. In some cases, this allowed a considerable amount of available energy (R_n − G) not counted in the partitioning of latent and sensitive heat flux (λE + H), which could cause significant discrepancies in the comparisons with the results from remote sensing. The errors inherent in R_n (net radiation), λE (latent heat flux), H (sensitive heat flux), and G (heat flux in the soil) were reported as 5–10%, 15–20%, 15–20%, and 20–30%, respectively, according to [28,29]. After closure, the λE values that represent the energy per unit area and per unit time were converted into evapotranspired depth unit for each time interval, resulting in actual evapotranspiration (ET_a), Equation (9).

$${\mathrm{E}\mathrm{T}}_{\mathrm{a}}=\mathrm{L}\mathrm{E}\mathrm{i}{\displaystyle \frac{3600}{\mathsf{\lambda }}}{\displaystyle \frac{{\mathrm{E}\mathrm{T}}_{\mathrm{r},\mathrm{d}}}{{\mathrm{E}\mathrm{T}}_{\mathrm{r},\mathrm{i}}}}$$

(9)

ET_a is the actual evapotranspiration (mm per day); ET_r is the reference evapotranspiration (mm); LE is the latent heat flux (w m⁻²); λ is the latent heat of vaporation (w m⁻²), subscripts d and i are for daily and instantaneous values, respectively.

The energy balance components and the ET_a estimated by the SETMI model were compared with the data measured in the field through the EC flux tower. To measure the accuracy of the model, RMSE (root mean squared error) and Bias indicators were applied according to Equation (10).

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}=\sqrt{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\displaystyle \frac{{({\mathrm{Y}}_{\mathrm{i}}-{\overline{\mathrm{Y}}}_{\mathrm{i}})}^2}{\mathrm{n}}}}$$

(10)

RMSE is the root mean squared error; n is the sample size; Y is the observed variable; $\overline{\mathrm{Y}}$ is the modeled variable.

3. Results and Discussion

Figure 2 presents the average daily temperature, daily gross precipitation and ETr over the study period 2008–2013 recorded for the automatic agrometeorological weather station.

Overall, the years 2012 and 2013 presented the lowest total precipitation rate with 298 mm and 291 mm, respectively. However, 2012 was considered the driest year due to the greatest accumulated atmosphere demand (ETr equal to 1959 mm) influenced by the highest average temperature equal to 12.4 °C, greater than the historical average for the region (11.0 °C). The year 2009 was the wettest year with accumulated precipitation of 633 mm, the lowest ETr (1404 mm) and average temperature (9.9 °C). Precipitation was very irregular over the study period, resulting in only 7% of the days with precipitation over 5 mm, distributed mainly from May to September (rainy season). Irrigation is required mainly after planting (by May), even though precipitation is higher from the planting date, it is common for dry spells to occur during the rainy season, so based on this analysis, irrigation is essential to maintain proper soil water content in the root zone for reaching high yields for corn and soybeans in Nebraska, mainly in the dry years such as in 2012 and 2013, intensifying the water conflicts over its use.

The energy balance components and ETa ratio estimated from the SETMI model based on TSEB agreed with the EC system as is shown in (Figure 3). The EB components value indicated acceptable performance with R² over 0.80 for all the variables. The RMSE showed low values, except for G with 45 W. m⁻². Generally, G values have little influence on energy balance because their values are considered very low, less than 5% of Rn in this study. So, the modeled results are consistent with ground flux measurements; as a result, ETa presented a high correlation between the values of the EC and modeled by the TSEB, overestimating the ETa values by 6% with R² of 0.96 and RMSE lower than 1.0 mm.d⁻¹ (0.73 mm.d⁻¹). Similar results were also observed by [29,30].

As shown in Figure 4 about the mean adjusted crop coefficient, when considering the average R² values across all the years (2008–2013) for all riparian vegetation, soybeans showed the greatest performance with a commendable R² value of 0.78. Following closely were corn and cottonwood, both yielding reasonable results with R² values of 0.67. Red cedar also demonstrated acceptable results with an R² value of 0.65. In contrast, bare soil and grass displayed comparatively poorer model fits, recording R² values of 0.46 and 0.42, respectively. The lower R² values observed for bare soil and short grass can be attributed to the pronounced variability in ETa across seasons. In the case of bare soil, the primary contributor to ETa is soil evaporation, which is significantly influenced by precipitation events. The values surge immediately following rainfall and decrease during dry soil conditions. This heightened sensitivity to precipitation events leads to greater variability in ETa for bare soil. A similar pattern is observed for short grasses in Nebraska. The combination of low canopy cover and rainfed conditions contributes to the high variability in ETa. The sparse vegetation cover and dependence on rainfall result in fluctuations in ETa, making it more challenging to establish a robust correlation with the chosen model. In contrast, for crops, the variability in Kc_adj is comparatively lower. This is particularly evident during periods of full crop cover when the impact of soil evaporation is minimized. During these phases, transpiration constitutes a significant portion (approximately 90%) of the ETa. The reduced influence of soil evaporation and the more stable transpiration contribute to a lower variability in Kc_adj for crops, leading to more reliable and consistent model outcomes.

Concerning the maximum Kc_adj values, it is noteworthy that soybeans and corn exhibited values of 0.9 and 0.8, respectively. According to the reference provided by [18], the anticipated maximum Kc_adj is 1.0 (based on tall grass reference). However, it is essential to note that in this study, water stress conditions were prevalent. This may be attributed to various factors such as rainfed conditions or suboptimal irrigation management across the seasons in many crop fields. Under these water stress conditions, the average regression curve for corn and soybeans exhibited lower values compared to the theoretical maximum of 1.0. The deviation from the expected maximum Kc_adj values suggests that these crops were likely experiencing limitations in water availability, leading to suboptimal conditions for transpiration and overall crop development. This emphasizes the importance of effective water management practices to ensure optimal crop performance and water use efficiency, especially in seasons prone to water stress such as the year 2012 and 2013. In the case of cottonwood and red cedar, the observed maximum Kc_adj values ranging between 0.70 and 0.75 suggest that these trees may not experience severe water stress because the presence of a deep root zone system in trees is a significant factor contributing to their ability to access water resources even in dry years. Additionally, they are in the riparian zone and there is a water table connected to the river. The deep root systems of cottonwood and red cedar enable them to tap into groundwater or other water sources at deeper depths, providing a more stable supply of water compared to shallow-rooted vegetation. This adaptability likely contributes to the observed moderate Kc_adj values, indicating that these three species are better equipped to withstand water stress conditions. Understanding the water use dynamics of different vegetation types, especially those with deep root systems, is crucial for effective water resource management and ecosystem assessment.

Even when irrigated crops like corn and soybeans pump from groundwater, they still interfere with surface water in riparian zones because groundwater and surface water are hydraulically connected. Pumping lowers the local water table, which can reduce baseflow to adjacent streams, alter stream–aquifer exchange, and diminish streamflow during dry periods. Moreover, return flows (percolation or runoff from irrigation) can modify the quantity and quality of water entering streams, influencing temperature, nutrient loads, and evapotranspiration patterns [5,31].

According to Figure 5 and

Table 2. Adjusted crop coefficient (Kc_adj) and accumulated actual evapotranspiration (ETa) for all the land cover for the period 2008–2013.

Land Cover	2008		2009		2010		2011		2012		2013		$\overline{\mathbf{x}}$
	ETa	Kcadj	ETa	Kcadj	ETa	Kcadj	ETa	Kcadj	ETa	Kcadj	ETa	Kcadj	ETa	Kcadj
Corn	714	0.64	639	0.68	600	0.56	623	0.60	797	0.58	821	0.66	699	0.62
Soybean	782	0.69	673	0.72	671	0.62	715	0.69	866	0.64	931	0.74	773	0.68
Short grass	539	0.48	604	0.65	613	0.57	584	0.57	535	0.40	621	0.51	583	0.53
Bare soil	407	0.37	519	0.56	480	0.44	525	0.51	394	0.29	399	0.32	454	0.42
Cottonwood	714	0.64	541	0.58	561	0.51	589	0.57	775	0.56	737	0.65	653	0.59
Red Cedar	648	0.58	558	0.60	542	0.50	568	0.55	620	0.45	729	0.58	611	0.54
$\overline{\mathrm{x}}$	634	0.57	589	0.63	577	0.53	600	0.58	664	0.49	706	0.58	629	0.56

Note(s): $\overline{\mathrm{x}}$: mean values.

, the Kc_adj varied over the years and the land cover influencing the accumulated ETa. In 2012, the lowest Kc_adj values were observed across all land covers, with a mean value of 0.49. This can be attributed to the exceptional drought conditions of that year, marked by elevated ETr due to high temperatures and low precipitation with poor distribution throughout the year. Under such circumstances, irrigation depths throughout the season may need to be higher to meet the crop water demands as seen in [4,32] for corn and soybeans. However, it is noteworthy that under conditions of elevated atmospheric demand (characterized by high ETr), the Kc tends to decrease. This phenomenon is exemplified by the trends observed in corn and soybeans in Nebraska, US, as documented by [33] and in tropical conditions [34]. This decrease is attributed to internal plant stomatal resistance, which limits the release of vapor from leaves into the atmosphere during periods of high atmospheric demand. In contrast, the year 2009, as the wettest year, had the highest Kc_adj (0.63). Conversely, the years 2013 and 2012, characterized by dry conditions, recorded the highest accumulated ETa values (706 mm and 664 mm, respectively). In such dry years, the pressure on water bodies such as rivers and groundwater becomes intense, potentially escalating conflicts over water use. This issue is particularly pronounced when farmers and public users from different states share the same water source, as exemplified by the Republican River in this case.

When considering each land cover individually, soybeans and corn exhibited the highest ETa values, recording 699 mm and 773 mm, respectively. This heightened ETa can be attributed to the greater Kc_adj values associated with these crops. The substantial water demand for soybeans and corn, as indicated by their higher Kc_adj, contributes to the increased evapotranspiration observed. On the other hand, bare soil displayed the lowest ETa value at 454 mm. The minimal vegetation cover and the absence of a transpiring crop contribute to the lower overall evapotranspiration for bare soil. Cottonwood, red cedar, and short grass presented intermediate ETa values. Cottonwood recorded 653 mm, red cedar had 611 mm, and short grass exhibited 583 mm. These values reflect the varying water use patterns of different land covers, with trees like cottonwood and red cedar demonstrating higher water use compared to short grass but lower than the intensive water demands of crops like soybeans and corn. In summary, the differences in ETa among land covers underscore the importance of understanding the water consumption patterns of various vegetation types, providing valuable insights for water resource management and land use planning. Indeed, understanding the water consumption patterns of riparian vegetation, especially in areas with pressure on freshwater resources from surface water, is crucial for addressing and mitigating water conflicts.

The high-water demand of riparian vegetation, as reflected in their evapotranspiration rates, can exacerbate water conflicts, particularly in regions where water resources are already stressed. Increased competition for water between agricultural activities, urban development, and riparian ecosystems can lead to conflicts over water allocation and use.

Effective water resource management strategies are essential to balance the needs of different stakeholders and ensure sustainable use of freshwater resources. This may involve implementing water conservation practices, optimizing irrigation techniques, and establishing clear regulations for water use in areas with riparian vegetation. In summary, recognizing the water consumption patterns of riparian vegetation is a key component of managing freshwater resources and mitigating conflicts over water use in regions where various stakeholders depend on the same water sources. In Nebraska, the Natural Resources Districts (https://www.nrdnet.org/, access date: 5 October 2025) provide a strong example of integrated groundwater–surface water management to protect riparian zones, demonstrating that linking irrigation permits to streamflow conditions and promoting coordinated management among users while also considering ET from riparian vegetation in decision-making, especially during dry years, can help safeguard river flows and reduce water conflicts.

Despite grass not having the highest Kc_adj and ETa compared to other land covers, short grass dominates as the primary land cover, constituting 50% of the total riparian vegetation area. Consequently, it also represents a significant contributor to water loss to the atmosphere. However, it’s noteworthy that most of the short grass in Nebraska is non-irrigated. As a result, ETa is primarily influenced by precipitation rather than drawing on freshwater resources from the ground or surface water (such as the Republican River basin). Corn and soybeans, together accounting for a substantial portion of the land cover at 15% and 3%, respectively, play a significant role. Given that most fields cultivating these crops are irrigated, both pumped groundwater and surface water directly impact the water source of the Republican River. Especially during dry years, these crops become significant users, underscoring their influence on water resources in the region. Cottonwood and red cedar, comprising 8% of the riparian vegetation, exert a relatively modest water demand. Despite their smaller proportion, this vegetation plays a crucial role in preserving water resources. Riparian vegetation, such as cottonwood and red cedar, plays a significant role in the water dynamics of riverine ecosystems contributing to stabilizing riverbanks and improving water quality.

4. Conclusions

Modeled energy balance components using SETMI showed a strong correlation to the ground data from EC, with ET presenting R² equal to 0.96 and RMSE of 0.73 mm.d⁻¹, affirming the model’s utility as a tool for assisting water management and decision-makers.

The Kc_adj values fluctuated over the years, impacting the accumulated ETa across different land covers. In 2012, characterized by exceptionally dry conditions, the lowest Kc_adj values were observed across all land covers, with a mean value of 0.49. Notably, soybeans and corn displayed the highest ETa values, registering 699 mm and 773 mm, respectively.

While grasses may not exhibit the highest Kc_adj and ETa compared to other land covers, their prevalence as the primary land cover, constituting 50% of the total riparian vegetation area, makes them a significant contributor to water loss to the atmosphere. It is important to note that most of the short grass in Nebraska is non-irrigated. Consequently, ETa is predominantly influenced by precipitation rather than tapping into freshwater resources from the ground or surface water, such as the Republican River.

Corn and soybeans, collectively representing 15% and 3% of the land cover, respectively, play a substantial role. Given that most fields cultivating these crops are irrigated, both pumped groundwater and surface water directly impact the water source of the Republican River. Particularly in dry years, these crops become major users, emphasizing their considerable influence on water resources in the region. Cottonwood and red cedar, comprising 8% of the riparian vegetation, exert a relatively modest water demand.

By providing consistent, transparent, and spatially explicit estimates of ET, remote sensing tools can help bridge disagreements over water availability and use them in transboundary basins. By incorporating these satellite-based ET datasets into compact compliance assessments and basin-scale hydrologic models, decision-makers can reduce uncertainty, enhance equity in water distribution, and strengthen adaptive responses to climate variability. Ultimately, remote sensing ET represents a critical tool for mitigating transboundary water conflicts by supplying a scientifically robust foundation for negotiation, monitoring, and governance in shared river basins. Integrating remote sensing–based evapotranspiration (ET) estimates into water management decisions can empower policymakers to better understand and manage water use in riparian zones, fostering more resilient and equitable strategies, especially where agricultural crops intensify competition for water resources. However, despite advances in technology, a key limitation of remote sensing remains the difficulty of estimating daily evapotranspiration (ET) at high spatial resolution; therefore, developing daily-scale ET models is essential to improve accuracy over large and heterogeneous regions, such as riparian vegetation zones.