Water Governance in an Era of Climate Change: A Model to Assess the Shifting Irrigation Demand and Its Effect on Water Management in the Western United States

Abstract

Communities throughout the United States have come to rely upon agriculture as a pillar of their political integrity, economic security, and health and wellbeing. Climatic conditions in the western portion of the United States necessitate most lands be irrigated to be arable. As a result, a major portion of the economy of the United States, and by extension the world economy, is driven by the continued viability of western United States water law and policy. Furthermore, due to the strong interrelationship between anthropogenic consumptive uses, streamflows, and wetland/riparian area ecology, irrigation demand has a strong effect on stream morphology, quality, and biology for aquatic species. Western water management is a complex mosaic that is controlled by western state, federal, and tribal governments. Each of these systems of law have vulnerabilities to climate change, which is well understood to cause increasing water supply scarcity. This articledemonstrates the risks climate change poses to our management of irrigation water demand, as well as the interrelationship between water supply and demand. Due to the shared nature of the resource, this article addresses both tribal reserved rights and state-based rights using data from Indian reservations that either contain and/or are closely adjacent to non-tribal agricultural communities. Those data are used in a systems–dynamics model to integrate crop–water requirement estimation techniques with climate change estimates and a Monte Carlo analysis to assess how irrigation demand could change because of changing temperature, precipitation, incoming radiation, and wind speed caused by climate change. Results indicate that climate change will cause increases in irrigation requirements at most locations. Further, climate change is expected to significantly increase seasonal variability in many locations. The model provides a useful tool based upon publicly available data that will allow individual water users to make conservation decisions necessary to preserve their water rights as the climate changes.

1. Introduction

There exists a complex mosaic of jurisdictional control over the land and water in the western United States. The region’s water law is dominated by two interrelated but distinct doctrines. First, the people indigenous to the region are entitled to reserved water rights pursuant to federal law for the maintenance of their homelands [1]. Named for the decision initially recognizing these rights—Winters v United States—the Winters doctrine guarantees that the United States holds in trust for the tribes water rights wherever “previously unappropriated waters are necessary to accomplish the purposes for which the reservation was [set aside]” [2].

Secondly, non-Indian water users acquire water rights based on the state law doctrine of prior appropriation. The legal predicate for acquiring a water right under this doctrine is the diversion of legally and physically available water and application to some beneficial use [3]. The amount of water a user is entitled to under this doctrine is likewise tied to beneficial use. In the case of irrigation, this amounts to the quantity of water “which, by careful management and use, is reasonably required [for] lands [during] such period of time as may be adequate to produce therefrom a maximum amount of crops as ordinarily are grown thereon” [4,5].

The prior appropriation and reserved water rights doctrines are linked by the West’s checkerboard of tribal and non-Indian lands, as well as the maxim “first in time is first in right”. Regardless of their legal basis, all water rights are administered together in priority based on the age of each water right [3]. Combined with the scarcity of water in the West, these factors often result in conflict between tribal and non-Indian water users. Although tribal irrigation water rights are often the most senior in a particular location, throughout much of the twentieth century, the United States federal government actively developed non-Indian water projects while simultaneously suppressing tribal development of water resources [6]. As a result, today tribal agriculture accounts for just 6.5% of the total land under cultivation in the United States [7]. Compounding matters, because they lack water infrastructure or use infrastructure that is in poor condition, actual water use invariably fails to track legal entitlement to water within tribal communities [8]. Although each reservation is unique, this sometimes results in water use that is just a fraction of non-Indians [9], while in other cases, tribes are forced to use out-of-date and inefficient irrigation techniques [10]. For differing reasons, these factors contribute to an overall lack of resilience to climate change within tribal communities.

Although the method for quantifying water rights for irrigation differs, all irrigation water rights are ultimately based on the amount of water necessary to grow the target crops given the region’s “soil conditions, method of conveyance, topography, and climate” [11]. The approach requires estimating current and future irrigation requirements based on historical conditions. Thus, the fundamental assumption underlying the quantification of irrigation rights is that water demand for crops in each location remains relatively static, allowing for the calculation of a quantity of water necessary for their irrigation. The continued viability of this approach is questionable in the era of climate change.

The effects of climate change on the water supply throughout the western United States has been thoroughly examined [12,13,14,15,16,17,18,19,20]. Much of that work has focused on assessing shifting snowpack trends, which is a primary driver of summertime water availability throughout the western United States [17,21,22,23,24,25,26,27]. Increasing variability in precipitation regimes and increasing trends in evapotranspiration across the western U.S. highlight the compounding effects of climate change on water supplies [28,29,30]. Climate change is increasing the likelihood of wildfires, which will also impact runoff dynamics and associated streamflow regimes [31,32,33]. Changes in streamflow due to climate change are variable across the western U.S. as a function of the hydrologic regime and model choice, while groundwater modeling shows declining water levels as a function of increasing evapotranspiration and groundwater pumping that exceeds recharge [34,35,36]. Assessing the various factors that will impact water supply in the western United States is critical for identifying potential adaptation strategies to support agricultural production into the future [37,38,39,40,41,42,43].

The impact of how climate change will affect the interrelationship between irrigation demand with decreasing water supply has been studied across scales, yet there are critical gaps in making those findings actionable with respect to western water law and management. Parameters that drive irrigation demand that have been examined in the context of climate change include temperature [30,44,45], precipitation [26,44,46,47], wind speed [30,44,48,49,50], and incoming solar radiation [30,51]. Associated impacts on crop yields, the amount of acreage under cultivation, and food production have been examined on the global, regional, and basin scales. Increases in crop water demand due to climate change have been shown at the global [52,53] and basin scales [54,55]. Researchers have estimated the effects of shifts in specific climate variables on irrigation demand within the growing season [52,53,56]. Many quantify the average annual irrigation demand [54,55] or seasonal/monthly irrigation demand [57,58] by projecting changes in evapotranspiration and precipitation through 2100. For example, Rajagopalan et al. projected both annual and seasonal irrigation demand for the northwestern United States’ Columbia River Basin based on future estimates of temperature and precipitation [59]. They found that changes in irrigation demand were sensitive to the crop type and concluded that temperature was a greater driver of the change in irrigation demand compared to shifting precipitation.

This study contributes to this body of research by developing a methodology for estimating changes in both monthly and annual irrigation demands due to climate change at the site scale. We estimate this by assessing not only the effects of changing temperature and precipitation but also dew point, wind speed, and incoming solar radiation using the input from the Multivariate Adaptive Constructed Analogs (MACAv2) over three time horizons: 2010–2039, 2040–2069, and 2070–2099 [44]. Moreover, rather than rely on a single estimate for each of these climate variables, we use a range of each variable, bounded by the 5% and 95% exceedance probabilities reported by MACAv2 projections for the RPC 8.5 scenario. The model uses system dynamics software to simultaneously estimate how these variables interrelate across their range of possible values to better understand the full range of potential irrigation demand outcomes. Based on those results, we explore the adaptability of western water law to the shifting water supply and irrigation demand and conclude with suggestions to mitigate its weaknesses and build more resilience into our water resource management.

2. Materials and Methods

This study focuses on 36 agricultural areas that exhibit heterogeneous land ownership (i.e., a mixture of tribal and non-Indian lands) and/or contain both tribal and non-Indian communities (Figure 1). We selected these areas for several reasons. First, we seek to bring awareness to the unique challenges climate change imposes upon tribal communities, which, despite experiencing some of the greatest adverse impacts from climate change, have largely been overlooked by the scientific community [14]. Equally important, there is a strong interrelationship between tribal and non-Indian water use throughout the western United States. For example, much of Arizona (sites 30 and 31), western New Mexico (sites 28 and 32), and eastern Oklahoma (sites 33–36) are nearly evenly split between tribal and non-Indian agricultural producers [7]. That trend also exists in certain parts of Nevada (sites 24), Montana (sites 4, 14, and 15), North Dakota (site 17) and South Dakota (sites 19 and 20) [7]. Other regions, such as southeastern California (site 29), northeast Washington (sites 2 and 3), northwest Montana (sites 7 and 8), and southern Nevada (site 26), have regions where up to 20% of all agricultural producers are tribal [7]. Each remaining site included in this study is located in a region where approximately 10% of the producers in the region are tribal [7]. Thus, by studying these particular agricultural communities, we are able to more accurately assess how climate change will affect the full suite of western water law and management—federal, state, and tribal. Finally, these sites cover a range of hydroclimatic zones to account for the complex geography of the western United States. The crop for each site was chosen to be the one most prevalent at that location based on 2023 United States Department of Agriculture cropland data [60]. Although irrigation is currently necessary at most of the study sites, areas that are currently dryland farmed were also included to estimate whether agriculture will continue to be viable in those locations without irrigation.

The irrigation demand model was developed using a mixture of data sources, either site-specific observed daily data from Agrimet and Mesonet, or for locations without observational data, data were obtained from the gridMET dataset [61] using Climate Engine [62]. The gridMET data resolution is 4 km² and includes the maximum temperature (°C), minimum temperature (°C), dewpoint temperature (°C), precipitation (mm), wind (ms⁻¹), and downward solar radiation (MJm⁻²). The period of record for the model, 1 January 2003 to 31 December 2022, was chosen because of data availability.

2.1. Model Development

The irrigation demand model presented here is based on the following soil–water balance equation:

$$\mathsf{\Delta }\mathrm{SW}={\mathrm{P}}_{\mathrm{e}}+{\mathrm{W}}_a-{\mathrm{ET}}_{\mathrm{c}}$$

(1)

where ΔSW is the change in the soil water content (mm), P_e is the effective precipitation (mm), W_a is the water applied to the soil (mm), and ET_c is the crop evapotranspiration (mm).

Crop evapotranspiration (ET_c, mm d⁻¹) was estimated using a crop coefficient technique, which included use of a reference ET (ET_r) modified using the dual K_c method for each location as follows:

$${\mathrm{ET}}_{\mathrm{c}}=({\mathrm{K}}_{\mathrm{cb}}{\mathrm{K}}_{\mathrm{s}}+{\mathrm{K}}_{\mathrm{e}}){\mathrm{ET}}_{\mathrm{r}}$$

(2)

where K_cb is the basal crop coefficient, K_s is the water stress coefficient, and K_e is the soil evaporation coefficient. Basal crop coefficients were estimated using techniques explained in Allen et al. (1998) [63] and elaborated in Allen et al. (2005) [64], with modifications noted in Annex 1 of Allen et al. (2007) [65]. The K_cb was modified with the water stress coefficient K_s because the actual ET is often lower because irrigation practices in arid western environs are often unable to maintain the ideal soil water content.

Reference evapotranspiration was estimated using the American Society of Civil Engineers standardized Evapotranspiration Equation [66]. That equation, which is based on the Penman–Monteith equation, is expressed as follows:

$${\mathrm{ET}}_{\mathrm{r}}=\frac{0.408\Delta \left({\mathrm{R}}_{\mathrm{n}}-\mathrm{G}\right)+\mathsf{\gamma }\frac{{\mathrm{C}}_{\mathrm{n}}}{\mathrm{T}+273}{\mathrm{u}}_2\left({\mathrm{e}}_{\mathrm{s}}-{\mathrm{e}}_{\mathrm{a}}\right)}{\Delta +\mathsf{\gamma }\left(1+{\mathrm{C}}_{\mathrm{d}}{\mathrm{u}}_2\right)}$$

(3)

where ET_r is the reference crop evapotranspiration (mm d⁻¹), R_n is the calculated net radiation at the crop surface (MJ m⁻² d⁻¹), G is the soil heat flux density at the soil surface (MJ m⁻² d⁻¹), T is the mean daily air temperature at a 1.5 to 2.5 m height (°C), u₂ is the mean daily wind speed at a 2 m height (m s⁻¹), e_s is the saturation vapor pressure at a 1.5 to 2.5 m height (kPa), e_a is the mean actual vapor pressure at a 1.5 to 2.5 m height (kPa), Δ is the slope of the saturation vapor pressure–temperature curve (kPa °C⁻¹), γ is the psychrometric constant (kPa °C⁻¹), C_n is a numerator constant (K mm s³ Mg⁻¹ d⁻¹), and C_d is the denominator constant (s m⁻¹). Units for the 0.408 coefficient are m² mm MJ⁻¹ [65,66].

The net irrigation requirement was estimated by determining the amount of water necessary to maintain the soil water content between the soil field capacity and plant water stress. The field capacity of a soil is defined as the point where the soil is holding the maximum amount of water that may be used by the plant [67], while agronomists commonly assume plant stress at around 50% to 60% of the field capacity [68]. Consistent with common irrigation practices, the model provides irrigation during the irrigation season once the modeled soil water content dropped to the estimated stress line. The model would then provide sufficient irrigation to bring the soil water content back to the field capacity using the following equation:

$${\mathrm{I}}_{\mathrm{n}}=\frac{\mathrm{ds}}{\mathrm{dt}}-{\mathrm{P}}_{\mathrm{e}}-\mathsf{\Delta }\mathrm{SW}+{\mathrm{ET}}_{\mathrm{c}}$$

(4)

where I_n is the net irrigation requirement (mm), P_e is the effective precipitation (mm), ΔSW is the change in the soil water content from the previous day (mm), ET_c is the crop evapotranspiration (mm), and $\frac{ds}{dt}$ is the difference between the field capacity and the current modeled soil water content (mm) [69].

The maximum daily irrigation was capped at 25.4 mm for sites where sprinkler irrigation is predominant, and depending on the crop, either 50.8 mm or 76.2 mm for sites in regions that predominantly flood irrigate. If the daily maximum irrigation did not provide sufficient water for the soil water content to rebound to the field capacity, irrigation would continue over additional days until field capacity was reached. Once the soil was at field capacity, the model would no longer provide additional irrigation water until the soil water content once again reached the estimated stress point.

2.2. Model Validation

The modeled ET_r using observed data from the Agrimet station at Kimberly, Idaho, was compared to the United States Bureau of Reclamation AgriMet model [70] and the University of Idaho’s ETIdaho model [71]. The AgriMet model results for Kimberly, Idaho, are available for the years 2003 through 2023, while ETIdaho results are available from 2003 through 2016. The modeled ET_c using observed data at Kimberly, Idaho, was compared to the University of Idaho’s ETIdaho model [71]. We also assessed model performance for estimating ET_r and ET_c when using gridMET synthesized input data by running the model validation a second time using gridMET input data in the place of the observed data for Kimberly, Idaho.

The modeled ΔSW was compared to output from the Washington State University (WSU) Irrigation Scheduler [68]. The WSU Irrigation Scheduler uses data from the Bureau of Reclamation’s AgriMet stations to estimate the annual soil water content using a soil water balance equation similar to Equation (1). For this study, we validated the irrigation demand model output for Kimberly, Idaho, using the Irrigation Scheduler’s 2014 soil water content estimates using the same soil characteristics, plant growth rate, and crop characteristics and assuming no irrigation took place.

2.3. Integration of Climate Change Scenarios

Climate change scenarios were obtained from projected climate data from the Multivariate Adaptive Constructed Analogs (MACAv2), which is a downscaling of outputs from twenty global climate models for two future climate scenarios [44,72]. The RCP4.5 (low emission) and RCP8.5 (high emission) scenarios correspond to radiative forcing levels of 8.5 Wm⁻² (~1370 ppm CO₂ eq.) and 4.5 Wm⁻² (~650 ppm CO₂ eq.) in the atmosphere by 2100 [73]. The RCP8.5 scenario was used in this study because current assessments indicate that greenhouse gas emissions continue to accelerate with little evidence of the collective political will necessary to transition to the RPC4.5 or similar scenario [74,75,76]. It assumes a steady rate of high greenhouse gas emissions as a result of trends in population and energy demand, coupled with modest rates of technological improvements [77].

Statistics for MACAv2 projections of seasonal climate variables (e.g., maximum temperature, minimum temperature, downward solar radiation, precipitation, and wind speed), were obtained for four time horizons: 1971–2000, 2010–2039, 2040–2069, and 2070–2099. Climate data for each location were used as the input to the irrigation demand model using Monte Carlo simulation. The percent change in each climate variable was calculated between 2000 (the base time) and each successive time horizon for both the spring and summer seasons at each location. The percent change for the 5% and 95% exceedance levels were the upper and lower bounds of the Monte Carlo analysis, where the potential change in irrigation at each time horizon was estimated through 500 simulations where the model randomly altered the climate variables between these bounds (Equation (5)) as follows:

$$x^{\prime }=x+\left|x\right|\ast m$$

(5)

where x′ is the daily value resulting from the Monte Carlo analysis; x is the daily data for each climate variable; and m is a multiplier, which is set as a random value between the lower and upper bounds for each time horizon. The results were used to estimate a future irrigation demand using Equation (2). Summary statistics of the future irrigation demand were then calculated for each location for each time horizon.

3. Results

3.1. Soil–Water Balance Model Validation

The modeled ET_r using observed input data had a strong correlation with the estimated results from both ETIdaho and the USBR Agrimet ET (R² = 0.99, Figure 2A,B) while the use of synthesized gridMET forcing data resulted in a weaker relationship between the modeled ET_r and ETIdaho and USBR Agrimet (R² = 0.89 and 0.88, respectively, Figure 2A,B). There were larger differences in the modeled ET_c (R² = 0.84 using observed data and 0.80 using gridMET data, Figure 2C). Variability could be attributed to differences in the k_cb curves used by each model, as well as assumptions for various ET model parameters regarding soil and plant characteristics, which affect both k_e and k_s. Monthly ET_r was closely aligned between USBR Agrimet, ETIdaho, and Modeled ET_r, while gridMET consistently predicted a higher ET_r (Figure 2D).

The irrigation demand model developed here provided soil water results that were comparable to the WSU Irrigation scheduler, capturing the observed soil water with an R² of 0.99 (Figure 3).

3.2. Model Results

The median change in annual irrigation across tribal lands in the western United States over the twenty-first century using the RCP8.5 scenario ranges from 154 mm to 331 mm, or 134–151% (Figure 4,

Table 1. Mean, median and standard deviation of the irrigation demand in each region over each modeling timeframe. Median and percent total change from the current value to 2099.

Statistic	Timeframe	Northwest	Northern Plains	Great Basin	Southern Plains	Southwest
Mean (mm)	Current	632	823	801	490	1069
	2039	687	846	853	683	1298
	2069	734	896	901	803	1370
	2099	793	953	979	883	1439
Median (mm)	Current	536	582	601	445	939
	2039	596	612	690	530	1163
	2069	645	768	793	578	1218
	2099	716	786	908	599	1270
	Change	180	204	307	154	331
	% Change	134	135	151	135	135
Standard Deviation (mm)	Current	6.2	7.3	10.5	8.1	10.5
	2039	7	7.9	10.9	8.5	12.8
	2069	7.2	8.3	11.2	8.9	13
	2099	7.3	8.6	11.3	9.4	13.3

). Although there is high spatial and temporal variability in annual irrigation volumes across the three time horizons, most locations indicate an overall increase in irrigation demand. The largest changes are predicted in the pasture land at site 17 in the northern Plains, winter wheat at site 29 in the Southwest, followed by sites 4 and 8 with alfalfa. Sites in the Northwest growing winter and spring wheat also showed high incidents of increased irrigation demand. Irrigation demand for alfalfa is expected to increase from 14–54% across the western US (Figure 4 and Figure 5). There is high variability within regions due to the heterogeneity of hydroclimates. For example, sites in the Northwest exist in some of the wettest parts of the continental United States (northwestern Washington), as well as some of the dryest parts of the region (central Washington and southern Idaho). Nonetheless, the irrigation demand is expected to increase for all locations.

Changes in monthly irrigation demand are variable within and across regions (Figure 6). While most sites show an increasing irrigation demand in most months, there are some locations that show a decreasing irrigation demand in a given month (e.g., June, site 25, Northern Plains, Figure 6B). Some locations show that the maximum irrigation demand will shift earlier in the year (e.g., site 3 in the Southern Plains shifts from June to May, Figure 6D). There are many locations that also show an irrigation demand when there historically had been little or no irrigation demand (Figure 6A–C). Many locations show high variability in the predicted monthly irrigation demand as a function of the variability in the climate data and simulation approach.

4. Discussion

Modeling site-specific changes in crop irrigation demand is critically important for irrigators and water resource managers alike. Here, we provide estimates of changes in crop irrigation demand over the next 80 years on western lands that are historically and economically important to both tribal and non-Indian agricultural communities (Figure 4, Table 1). This is important not only because indigenous people are widely acknowledged to experience the impacts of climate change in unique and significant ways [14] but also because the strong interrelationship between tribal and non-Indian water use necessitates that adaptation techniques be likewise integrated in order to be durable in the face of climate change.

This model incorporates a suite of climate variables important to accurately estimate the irrigation demand for the primary crops grown in each area [57,58]. The estimation of irrigation demand across the range of possible exceedance probabilities for these climate variables (ranging from 5% to 95% of MACAv2 projections) allows water users and managers to understand the full range of potential irrigation demand outcomes. Assessing the monthly irrigation demand, in addition to the annual demand, is important for understanding the variability in how the irrigation demand will shift in different regions and across cropping choices (Figure 6) [54,55]. Increased irrigation demand is not evenly spread across the entire irrigation season, and months with the largest change in irrigation demand are variable even within a given region (Figure 6). Early season changes in irrigation demand when crop stress is more detrimental to crop yield, quality, and profitability [78,79,80,81,82], and in the late season when water supply is most scarce are of particular concern. Therefore, uncovering when within an irrigation season the irrigation demand will experience the greatest change is critically important. Finally, the model’s reliance on publicly available data and toolsets make it broadly accessible, while use of site-specific data make the results directly relevant to individual decision-makers in comparison to existing regional models. The model is useful for individual water users and water managers alike, as well as legal and policy decision makers, as they all grapple with the shifting irrigation demand stemming from climate change.

4.1. Model Utility for Water Users

This model will help individual water users adapt to climate change by providing site-specific estimates of long-term irrigation water demand across the western United States, which would allow them to develop site-specific long-term plans to acquire an additional water supply, either through the acquisition of additional water rights or the integration of conservation measures into their irrigation practices. For example, a water user growing winter wheat near site 11 in Oregon can see that the irrigation demand is likely to increase by 66 mm in the short term, by about 117 mm by the middle of the century, and by 260 mm by 2099 (Figure 5). This information allows water users to develop a plan for acquiring additional water supplies, for assessing the viability of existing crop choices, and/or implementing increased efficiency standards. Such a plan could be executed over a time horizon that suits individual water users, thereby allowing them to absorb those costs over that time window, which would increase the likelihood of successful adaptation.

The model is particularly useful for junior water users, who will face increasing pressure to curtail their use as water demand for senior users grows and water supplies continue to shrink from climate change. Historically, junior users have often negotiated to reduce their long-term use in exchange for safe harbor when a shortage occurs. For example, a key component of a recent settlement on Idaho’s Eastern Snake Plain protected junior groundwater users from future water calls in exchange for reducing their water consumption and providing mitigation water to the senior users in the basin. That reduction in use was shared collectively by thousands of groundwater users, notwithstanding their individual priority dates [83]. Similar arrangements exist elsewhere in the West, each fashioned to address the unique needs and characteristics of the water user community. Rather than wait for a shortage, the model described herein could help junior water user groups to forecast future irrigation water demand and engage in prophylactic measures to reduce the likelihood of a water call or negotiate safe harbor before a shortage occurs.

4.2. Usefulness for Broader Water Adaptation Policy

The model is useful to decision-makers as they seek to update the water laws of the western United States in a way that helps to adapt to climate change. For example, results indicate that irrigation demand is expected to increase by 34–51% in the Southern Plains (Figure 4). However, much of this region is currently dryland farmed, resulting in a water law regime that is less developed compared to the traditionally arid states. These results indicate that irrigation will become increasingly common in this region and policies will need to be implemented to usher in that new era in a coherent way. In addition, as the water supply decreases, policy-makers in this region will need to implement increasingly stringent policies related to water use. As we discuss in the next section, many tools already exist but have historically been applied by the judiciary in a way that blunts their effectiveness, particularly as climate change accelerates water demand and water shortages [84]. Policy-makers within the legislative branch have wide latitude to fix these judicial limitations through the development of stricter water laws, as well as policies that support stricter application of the laws already on the books. The broadly relevant results of this model should provide a catalyst to encourage policy-makers to control water appropriation and use more aggressively.

The model is also useful to the judiciary as it seeks to update its water law jurisprudence to adapt to climate change. A major impediment to adaptation has been the courts’ expansion of the “property right” aspect of water rights, subjecting state governments to liability for just compensation should a court determine that a particular water management decision resulted in the taking of someone’s water right [3]. As a result, western water agencies are often reticent to strictly enforce water laws that might curtail inefficient or unreasonable use [84]. A fundamental precept within the prior appropriation doctrine, however, is the maxim that the property right one holds in the use of water has never been absolute. Instead, water rights are expressly conditioned on the continued reasonable use of the water and do not provide a property right to waste water or use it in a grossly inefficient manner. Thus, models such as this one provide useful results that may encourage the judiciary to move back toward the center of the prior appropriation doctrine and empower regulators to more aggressively control water appropriation and use.

4.3. Usefulness for Water Management Decisions

In addition to the broad policy considerations outlined above, the model—coupled with watershed-scale models related to water availability, remote sensing-based methods for assessing water consumption, and other decision support tools—will be useful for water managers as they address the increasing prevalence of water management decisions that are affected by climate change.

Western water law currently has several tools at its disposal to provide for both long- and short-term techniques to adapt to climate change. In the short term, the site- and region-specific information provided by this model could help water managers to determine when and where to more rigorously apply legal doctrines, such as requiring new water rights be in the public interest and existing rights be limited to their reasonable need [3,4].

Although the scope of the public interest doctrine differs from state to state, it generally places the burden on the water right applicant to demonstrate that the new water use will not adversely affect fish, wildlife, and other aquatic habitats; recreation; aesthetic beauty; transportation; navigation; water quality; and other important economic and ecologic values within the community [85,86]. Considerations related to climate change fit naturally into the underlying rationale for this rule. The model developed herein is a useful tool for water managers, who would be charged with assessing water availability both now and into the future with associated climate change impacts in a particular basin. If the results indicate that as the climate changes a new water right will eventually adversely affect the public interest, then those implementing agencies should be required to place conditions on the water right or deny it outright. Relatedly, many states require water to be physically available before a new water right may be issued [87,88]. However, as irrigation demand increases from climate change, the water currently available might be required to serve more senior water rights. This model could help water managers assess not only current water availability but also assess future water availability as a function of increasing water demand, thereby potentially reducing future conflict when increasing demand outstrips the supply. Likewise, the information provided by the model would be useful to the judiciary when called upon to review the decisions made by water managers. If applied thoroughly, the approach would act as a brake, helping communities from further over appropriating water supplies that promise to continue to decline with climate change. Although these laws do not apply to use pursuant to federal law by the United States and tribes, these sovereigns could more broadly implement similar concepts as a catalyst for adaptive governance within their own jurisdictional spheres so as to help ensure increased water security moving forward.

The model could also provide useful information to water managers as they assess the ongoing reasonableness of individual water uses within their jurisdiction. The state law concept of reasonable beneficial use is particularly crucial in this effort. That doctrine limits a water right to the amount reasonably necessary for a particular beneficial use, which is estimated based on the region’s “soil conditions, method of conveyance, topography, and climate” [11]. The mirror image of reasonable use is prior appropriation’s prohibition of the waste of water, which provides that water used in an unreasonably inefficient manner may be forfeited without causing a taking [3,84]. Thus, by providing water managers with an idea of how irrigation demand will change with time, this model provides a useful way for those managers to assess the ongoing reasonableness of a given individual’s water use. Moreover, because it can be used to estimate regional increases in irrigation demand, this model can be coupled with climate–water supply models to provide water managers with a better idea of the future interrelationship between water supply and demand in a given location. Armed with those data, these managers could then implement phased-in increased water efficiency standards, other conservation efforts, or water supply development programs in an effort to ensure that the water supply continues to match the demand as climate change progresses. Because of their sovereign status and because their rights exist pursuant to federal law, reserved water rights held for the benefit of tribes are not subject to these state law concepts. However, consistent with their own values regarding water use, tribes are free to promulgate tribal water codes that incorporate these or similar concepts as they allocate water within their jurisdictions.

Certainly, strict water efficiency standards can sometimes result in unintended consequences, particularly in regions where downstream appropriators have come to rely on someone else’s return water [89]. Recognizing this principle, efficiency standards should move beyond solely considering individual efficiency and instead consider watershed-scale water balances. Regardless, generalized notions of unintended consequences should not paralyze us as we continue to unravel the complex web that is water systems engineering in many basins across the West. Ultimately, western water law does not protect demands from subsequent users where someone continues to waste water for their benefit [90], and the single best way to ensure that the water supply can meet the demand of existing and new uses is through conservation.

To be successful, the reforms above must be coupled with governmental incentives to accelerate investments providing for water efficiency and conservation, watershed-scale modeling, and offset planning and implementation costs to improve the western water infrastructure necessary to adapt to climate change (e.g., Inflation Reduction Act [91]). The model would be useful in this regard as well, helping policy-makers to prioritize regions with the highest need so as to efficiently allocate limited resources and blunt the harshness of the water law reforms suggested herein. For example, the model results indicate that irrigation demand near site 21 in northwestern Washington is expected to increase substantially while changes are expected to be more modest in the center and eastern sides of the state. Based on this information, water managers might choose northwestern Washington for management prioritization and policy-makers could likewise provide increased resources for that effort while also developing policy and investment to support the region’s climate adaptation. Undoubtedly, other considerations such as the shifting water availability, ecological concerns, and economic output also drive decisions regarding the allocation of finite resources. However, the information provided by this model regarding the increasing water demand would be useful as policy-makers and water managers engage in the calculus required to prioritize limited resources for both water regulation and investment in climate resilience across jurisdictions.

Although beyond the scope of this article, myriad approaches have been adopted ad hoc throughout the United States that could be scaled to the policy level for more meaningful adaptation to climate change. Policy-makers could continue to invest in watershed-scale models that are developed as decision support tools and develop tax incentives that would encourage individual water users to invest in more efficient infrastructure, as well as direct grant funding to subsidize the costs associated with water infrastructure improvements (e.g., NRCS Regional Conservation Partnership Program). Similarly, government funding could be invested in buying out water uses on a seasonal or permanent basis (e.g., conservation easements). Likewise, state and tribal governments could continue to develop water banks whereby seniors have the option of temporarily reducing their use and market the excess water, which allows other water users to adjust to short-term variability in a way that requires less systematic redesigns in infrastructure [92]. Alternatively, those that need access to additional water supplies in water-short regions could subsidize infrastructure upgrading costs that would allow senior users to improve their efficiency, thereby freeing up the water supply for use by others [93]. Although precautions to protect existing users must remain as we broaden our use of these techniques, recalibrating the relative value placed on system-wide benefits versus individual rights would effectively streamline water markets and transfers, which would incentivize private water conservation efforts and provide for more efficient use of the water resource [89].

Finally, although these incentives would undoubtedly be effective at the farm scale, significant direct governmental investment will likely be necessary to reinvent our broader water resource infrastructure to build the necessary resilience to climate change. Only the United States federal government has the resources to make that type of large-scale investment, and this model could be useful for prioritizing the regions in the greatest need. Importantly, the federal government has long been involved in western water resource management, primarily through federal development and management of dams, reservoirs, and other water infrastructure [94,95,96]. Although the twentieth century dam building era has come to a close, the federal government could pivot its resources and expertise into a twenty-first century era focusing on water conservation practices and associated improvements to our monitoring infrastructure. To make this a reality, tribal and state governments could partner with the federal government to usher in a new era in federal water resource development, which would focus on retrofitting the existing water infrastructure and developing new water projects in order to monitor and use water as efficiently as technologically and economically feasible. This is occurring in an ad hoc way in many western watersheds using funding from entities such as the USDA and USBR. These efforts are often led by local decision-makers that could benefit from additional strategic planning resources that would facilitate collaboration and coordination between relevant entities.

5. Conclusions

The model results indicate that the irrigation demand in the western United States is generally going to increase on both an annual and seasonal basis and/or become increasingly variable as climate change continues. As a result of its reliance on site-specific public data to estimate the increasing irrigation demand over three time horizons, the model provides a useful tool to help individuals make effective phased-in conservation goals in the face of these changes. However, the increasing water demand will be coupled with water availability that is expected to become increasingly scarce, mistimed, and variable. As a result, without intervention and adaptation, climate change will likely result in increasing conflict and decreasing certainty within the western United States over the next century.

The model also provides a useful tool as we grapple with the broader water policy necessary to adjust to climate change. Many aspects of western water law are not well positioned to adapt to the changes that will likely occur to water supply and demand because of climate change. The most substantial step water users could take to reduce their shared risk and uncertainty would be to invest their significant political capital toward curtailing climate change. Short of this, water users and western lawmakers should focus on adapting to changing conditions by reforming water law and policy to rebalance the relative value we place on water efficiency versus property rights. The practical, political, and economic reality of such a shift necessitate that it coincides, however, with policies that support broad improvements to our water resource infrastructure, as well as incentivize decisions that maximize water flexibility and efficiency. Although several promising approaches are suggested herein, specific adaptive management will be driven by the unique circumstances in each community. Generally, however, techniques that integrate flexibility, incentivize efficiency, and provide a safe harbor from curtailment on either a permanent, long-term, or seasonal basis seem to hold the most promise [97].

Finally, cooperation between the sovereigns—the states, tribes, and federal government—will be crucial as each works within its jurisdiction to mitigate the effects of climate change. The federal government has a unique role as the only entity with sufficient resources to support many of the system-wide infrastructure measures and mitigation techniques that will become necessary. Further, to the extent that hydrologic systems transect political boundaries, lawmakers should explore cooperative approaches that address water resource management from a wholistic, basin-wide perspective. Ultimately, the challenges brought by climate change require that governance at all levels abandon stubbornly rigid and outdated models and adapt into new modes that rely on flexibility, creativity, and cooperation.