United States v. Gila Valley Irrigation District

*1447 PHASE IV MEMORANDUM AND ORDER — AMENDED

This is the fourth phase of the most recent round of litigation concerning the enforcement of the Globe Equity Consent Decree of June 29, 1935 (hereafter “Decree”) 1 for the benefit of the Gila River Indian Community (hereafter “GRIC”) and the San Carlos Apache Tribe (hereafter “Apache Tribe” or “Apaches”). The Court held trial November 7-17,1994 and heard closing arguments January 18, 1995. The Court heard evidence and argument related to the following issues:

(1) whether the Apache Tribe’s rights to the natural flow of the Gila River are compromised by farming activities in the upper valleys;

(2) whether apportionments to the Upper Valley Defendants (hereafter “UVDs”) are subject to prior calls of both the San Carlos Apaches and the GRIC;

(3) whether Article VTII(4) requires that there be sufficient water stored for the proper irrigation of 80,000 acres of San Carlos Project lands before any apportionment may be made to UVDs;

(4) whether the Decree prohibits the diversion of the entire flow of the Gila River by defendants in the DuncanVirden Valley pursuant to the “Cos-per’s Crossing condition”;

(5) whether, under Article VIII(5), UVDs are prohibited from diverting an available apportionment unless there is water presently stored and available for release in the San Carlos Reservoir;

(6) what method the Water Commissioner should employ in deducting transit and seepage losses;

(7) whether the Decree endows the lands in the Gila Crossing District with priority rights that may be enforced as against all upstream parties to the Decree;

(8) what reporting scheme the Water Commissioner should adopt for ensuring that water is not diverted for use on lands not “then being irrigated,” and whether the “then being irrigated” clause and the reporting scheme apply equally in the lower valleys;

(9) whether the Decree prohibits the use of the waters of the Gila River for purposes other than irrigating “crops of value”;

(10) whether the call system being implemented for managing the Gila River complies with the legal requirements of the Decree; and

(11) whether the Court should order the permanent removal from the Decree those lands that have been determined to be urbanized by the Arizona Department of Water Resources.

It is unnecessary to restate the facts here, as they have been carefully laid out in prior decisions of this Court and the court of appeals. See, e.g., United States v. Gila Valley Irrigation Dist., 31 F.3d 1428 (9th Cir.1994); United States v. Gila Valley Irrigation Dist., 961 F.2d 1432 (9th Cir.1992); United States v. Gila Valley Irrigation Dist., 454 F.2d 219 *1448 (9th Cir.1972); United States v. Gila Valley Irrigation Dist., 804 F.Supp. 1 (D.Ariz.1992). The Court’s findings of fact relative to the specific issues decided here are set forth as necessary below.

I. WATER QUALITY

The Apache Tribe argues that the waters of the Gila River are contaminated by farming practices in the upper valleys and that the Tribe is therefore deprived of its right under the Decree to the “natural flow” of the stream. UVDs argue that the water quality deteriorates as the river passes through the upper valleys due to natural causes, some of them unknown. For the reasons expressed below, the Court finds (1) that the water reaching the San Carlos Reservation is of significantly lower quality than the water in the upper valleys, (2) that UVDs are responsible to a significant degree for the degradation of the water reaching the reservation boundary, and (3) that equity requires that certain measures be adopted to protect the Tribe’s water right.

The courts of the western states generally agree that a prior appropriator of water is entitled to protection, including injunctive relief, against material degradation of the quality of the water by junior appropriators upstream. See, e.g., Phoenix Water Co. v. Fletcher, 23 Cal. 481, 487 (1863); Salt Lake City v. Boundary Springs Water Users Ass’n, 2 Utah 2d 141, 270 P.2d 453 (1954). Further, the Supreme Court has held:

What diminution of quantity, or deterioration of quality, will constitute an invasion of the rights of the first appropriator will depend on the special circumstances of each case, considered with reference to the uses to which the water is applied. A slight deterioration in quality might render the water unfit for drink or domestic purposes, whilst it would not sensibly impair its value for mining or irrigation. In all controversies, therefore, between him and parties subsequently claiming the water, the question for determination is necessarily whether his use and enjoyment of the water to the extent of his original appropriation have been impaired by the acts of the defendant.

Atchison v. Peterson, 87 U.S. 507, 514-15, 22 L.Ed. 414 (1874). See also Arizona Copper Co. v. Gillespie, 230 U.S. 46, 56-58, 33 S.Ct. 1004, 1006-07, 57 L.Ed. 1384 (1913). 2 There is no substantial dispute that the United States and the Apache Tribe bear the burden of proof on this issue. E.g., Pacific Portland Cement Co. v. Food Mach. & Chem. Co., 178 F.2d 541, 546 (9th Cir.1949). 3

A. Overview

The Gila River once supported irrigation from its surface flow in regions extending from above the New Mexico border to the confluence of the Gila and the Salt River. The river is now overdeveloped and overallocated. In the upper valleys, surface flow is heavily augmented with water pumped from wells. Further, the growers in the upper valleys on occasion divert the entire flow of the stream into irrigation canals to serve the acreage they farm. The return flows from diversions are often recycled, diverted again and applied to other fields. The effects of these and other practices contribute to a dramatic increase in both the salt load and the salinity of the Gila River. The salt load is increased, in part, by pumping groundwater that is higher in salt content than the *1449 surface flow and mixing it with the stream. Salinity increases as the river flows through the Salford Valley due to the diversion of the entire stream and the consumptive use of the water, but not the salt, by irrigated crops and other plants. By the time the flow, if any remains, reaches the San Carlos Reservation boundary, the water is degraded to such a degree that the Apache Tribe may be unable to grow the types of vegetable and grain crops that were once raised there.

B. Water Quality of the Gila River

There are two interrelated aspects to the water quality issue in this case: flow rate and contamination. The contaminant is salt, which prevents the successful cultivation of a variety of crops. 4 The issue of flow rate is, in this case, inseparable from the issue of salt, given the general agreement among the testifying experts that flow rate is a primary factor affecting water quality: at a high level of generalization, the lower the flow rate, the poorer the quality of the water. 5

To see the difference in the quality of the water available to the UVDs and that available to the Apache Tribe, one need only compare salinity in the upper valleys and salinity at the reservation boundary. Plaintiffs’ expert witnesses conducted extensive field studies in 1993 and 1994, sampling water and measuring flow rates at various sites in the upper valleys and on the reservation. Their data for June 27 and 28, 1994, for example, revealed total dissolved solids (TDS) of 600 milligrams per liter (mg/l) at the head of Safford Valley (below the Brown Canal diversion point). 6 In the middle portion of the valley, above Smithville Canal, the TDS increased to 1,250 mg/1. At the lower end of the valley, above the Fort Thomas Canal diversion, TDS was 1,382 mg/l. And at the reservation boundary, TDS was measured at 1,992 mg/1. See Stetson Engineers, Inc., Preliminary Review of the Suitability of the Quality of Gila River Water to Irrigate Lands of the San Carlos Apache Reservation, Arizona 1-34 (September 15, 1994) (hereafter “Stetson Water Quality Report”). Further evidence related to these field studies was submitted at trial. Samples taken between August 17 and 19, 1994 revealed a similar trend, although the salinity was even greater in the lower reaches of the upper valleys. Specifically, the engineers measured TDS of 657 mg/l at Brown Canal, 1,638 mg/l at Smithville Canal, 1,690 mg/l at Fort Thomas, and 1,830 mg/l at the reservation boundary. 7

The evidence at trial also demonstrated unequivocally that flow rate is inversely related to salinity. Plaintiffs’ experts employed data collected by John D. Hem 8 in *1450 the 1940s to prepare graphic depictions of the relationship between flow rate and electrical conductivity (EC) as the river flows downstream from Blue Creek, New Mexico, to Calva, Arizona. The evidence shows that as the flow rate decreases in the Safford Valley, EC rises. In fact, the difference can be quite dramatic, with EC levels often as low as 1,000 microSiemens per centimeter (cS/cm) in the central reach of Safford Valley and as high as 5,000 cS/cm on the reservation. See Stetson Water Quality Report at 1-22,11-12 to 11 — 14, Table II-4 & Figure II-3. In graphical representations, the rapid divergence of lines representing EC and flow rate is apparent. EC begins to rise at the head of Safford Valley, while flow rate declines. Salinity peaks in the neighborhood of Fort Thomas, about 25 miles downstream from Safford. Beyond Fort Thomas, at the bottom of the Safford Valley, the water quality stabilizes or improves. See generally Stetson Water Quality Report 1-20 to 1-25 (analyzing Hem’s data). Precisely the same trend is reflected in United States Geological Survey (USGS) data and the Stetson field data collected in 1993 and 1994. See id. at I-36, Figure I-10b. On the other hand, if the flow is maintained at or above 100 cfs, EC remains stable at values well under 2,000 cS/cm. Id. at 1-23-24, Table 1-6 & Figure 1-6.

C. Causes of Degradation of Water Quality

Plaintiffs, the government and the Apache Tribe, argue that the degradation of the Gila River water quality is the direct result of various fanning practices in the upper valleys. UVDs dispute this. They argue that the degradation, if there is any, is the result of purely natural causes and that the water would be similarly degraded if cultivation in the upper valleys ceased and the lands were taken over by wild vegetation.

The Court finds abundant evidence supporting plaintiffs’ view. There are two primary causes of the increased salinity and salt load in the river, pumping from the groundwater and diverting the entire natural flow of the stream.

Pumping from groundwater: The effect of pumping on the salt load and salinity is difficult to overstate. The amount of water pumped prior to the entry of the Decree was apparently quite small. By 1938, there were about 30 wells in the upper valleys. See Gookin Engineers, A Technical Report on Various Aspects of Globe Equity No. 59, ch. 1, at 15 (August 1994) (hereafter “Gookin Technical Report”). In each of the years 1937, ’38, and ’39, about 20,000 acre feet (AF) were pumped from the groundwater. Trial Transcript (hereafter “Tr.”) at 230-32 (November 8, 1994) See also Stetson Water Quality Report II-4 to II — 6, Table II — 1 (showing pumping in years after 1940). The number of wells and the amount of water pumped steadily increased. Gookin Engineers estimates that there were 900 wells in the Safford Valley by 1958. Gookin Technical Report, eh. 1, at 16. In recent years, the amount of groundwater pumped in the Safford Valley has varied between 214,100 AF (1977) and 62,400 AF (1986). Stetson Water Quality Report II-5. In general, UVDs pump greater amounts of water in dry years, when the surface flow is low, and lesser amounts in wet years, when there is more surface flow to meet their irrigation needs.

The expert witnesses disagreed about the extent to which groundwater pumping affects the water quality. The greater weight of the evidence and the more plausible analysis support plaintiffs’ argument that the effects are significant. First, some of the wells in the Safford Valley tap into very high salinity water sources. The evidence was uncontroverted that there are natural high-salinity artesian seeps flowing into the Gila, but the flow from these seeps is very low, so their effect on overall salinity is minimal. Certain “hot” wells, however, are responsible for localized TDS elevations of several hundred percent. 9

Second, all of the evidence indicates that the groundwater is higher in salts than undiverted surface flows. Pumping from the groundwater adds salt to the surface flow of the Gila River. The increased salt load in *1451 pumped water contributes directly to the increase in salinity as the river flows through the Safford Valley. Moreover, pumping during the dry summer months places the lower quality water in the stream at the time when natural flows are at their lowest and demand on the stream is at its highest.

Third, groundwater pumping contributes to a decline in the water table, leading to a “losing stream” phenomenon, in which the river loses water to the groundwater. Surface flow may be seriously depleted or eliminated due to nearby pumping. Also, a losing stream condition can exacerbate the salinity problem to the extent that surface flows of higher quality, such as from precipitation and sudden freshets, go to recharging the alluvial aquifer, rather than remaining on the surface.

Fourth, excessive pumping alters the relationship between the alluvial aquifer, immediately underlying the riverbed, and the basin-fill aquifer, which is an older geologic formation up to a hundred feet below the surface. Water in the basin-fill aquifer is at a higher pressure than that in the alluvial aquifer. When large amounts of water are pumped from the alluvium, the piezometric relationship between these two aquifers is altered and the water from the basin-fill aquifer seeps upward into the alluvial aquifer. Because the basin-fill aquifer is much higher in salts than the alluvial aquifer, this seepage increases salts in the alluvium and in the river.

Diversion of the entire flow of the stream: The evidence demonstrates that UVDs on occasion divert the entire flow of the stream at various canal headings. The water arriving back at the river consists largely of irrigation return flows that have absorbed salts from the soils to which the water was applied. This cycle is repeated when the entire flow is diverted at canal headings downstream. Thus, the river becomes an “agricultural drain” carrying salts from the soils of the cultivated acreage. Also, diversion of the entire flow of the river exacerbates the deleterious conditions created by pumping, including the losing stream phenomenon and the upward seepage from the basin-fill aquifer into the alluvial aquifer. 10

These practices contribute to salt load and salinity in a fairly straightforward manner: As water is diverted and used for irrigation, the ratio of salt to water increases, because the plants consume only the water, leaving the salt behind. Additions of salt to the system, from both natural and artificial sources, add to the burden of salt in the stream.

UVDs acknowledge that the quality of the Gila River deteriorates as it proceeds through the Safford Valley toward the San Carlos Reservation and that this deterioration is attributable to increased concentrations of salts. Further, UVDs concede that the greatest contributor of salt in the Safford Valley is the basin-fill aquifer. UVDs argue, however, that the influx of salt in the Safford Valley is from a natural cause. They further argue that the water is not significantly worse than it was in the early 1940s and that the Apaches would receive equally salty water if irrigated agriculture in the Safford Valley ceased.

The Court cannot agree that the influx of salt in the Safford Valley is primarily natural. The parties agree that the salt load increases significantly in the vicinity of Fort Thomas. UVDs’ expert witness, Dr. Gerald Matlock, states that this inflow of salt is attributable to (1) the natural upward migra *1452 tion of saline water from the basin-fill aquifer, and (2) a geologic phenomenon he refers to as the “Fort Thomas anomaly.” The Court agrees with plaintiffs, however, that Dr. Matlock overstates the effect of natural migration from the basin fill to the alluvium. 11 Nor is the Fort Thomas anomaly a plausible explanation for the salt. Dr. Mat-lock hypothesizes that there are unknown geological forces at work creating an enormous salt influx from the basin-fill aquifer in the area of Fort Thomas. Dr. Matlock’s argument is speculative and lacks support in the form of physical evidence. It is, moreover, directly at odds with the more carefully documented conclusion of other expert witnesses that, absent pumping from the groundwater, the natural migration of water from the basin-fill aquifer has a relatively small effect. Pumping, however, causes immediate and significant increases in upward flow from the basin fill, in addition to directly affecting salinity by bringing the saline groundwater to the surface.

UVDs also argue that the quality of the water leaving the Safford Valley has not changed in over fifty years. UVDs rely almost exclusively on a trial exhibit comparing EC and flow rate in the Safford Valley in July of 1941 and August of 1994. The exhibit indicates that EC in 1994 was about the same, if not somewhat improved, relative to EC in 1941. The Court agrees with plaintiffs that the exhibit does not compare two similar moments in the river’s history and provides little valuable information concerning the long-term degradation of the river. First, the conditions that existed in 1941 are different from those of 1994. Three very dry years preceded 1941. 12 The evidence at trial indicated that flows will be more saline during extended periods of dry weather or drought because (1) when flow is diminished the concentration of salts is higher and (2) when flows are low for an extended period, there is a corresponding lack of “flushing” of the salts from the riverbed and soils. The high flows in 1941 would, according to the testimony, be expected to flush out the salt buildup. In marked contrast, 1994 followed on the heels of three of the wettest years this century. 13 These extremely high flows would likely have left the riverbed relatively free of salt and the alluvial aquifer recharged with low-salinity water. Second, UVDs erroneously assert that the exhibit demonstrates that pumping is not the cause of the increased salinity. In fact, UVDs were pumping 20,000 to 25,000 AF of water for five years prior to 1941. See Tr. at 230-31 (November 8,1994). Although pumping has significantly increased since that time, it cannot be plausibly asserted that pumping would have had no impact on the river system prior to 1941. Finally, in 1994, the Water Commissioner, acting pursuant to this Court’s order, maintained a live stream condition in the Safford Valley. This could have had a measurable impact on salinity. For these reasons, 1941 and 1994 make a poor comparison of the long-term condition of the river.

UVDs also argue that the Apache Tribe would receive equally degraded water if no irrigation at all occurred in the upper valleys, because UVDs’ lands would be taken over by phreatophytes. 14 UVDs argue that their *1453 crops consume only about 2.5 acre feet per acre (AFA) per year and that if phreatophytes invaded their lands, the consumptive use would rocket to 5 AFA per year, leaving the river as salty, if not saltier, than it is now. UVDs assume, first, that half of their lands will give over to phreatophitie growth if crop irrigation ceased. There is no evidence to support this assumption. The argument also assumes, without supporting evidence, that wild phreatophytes will draw on the river in the same manner as a crop being irrigated during a growing season with diverted and pumped water. The evidence at trial indicates that this is an unreasonable assumption, both because wild phreatophytes do not receive water in the same way as irrigated crops and because UVDs would not, in any case, pump groundwater and divert the river to serve wild phreatophytes on their lands, some of which are several miles downstream of the canal headings.

Based on the evidence discussed here and the totality of the evidence presented at trial, the Court finds that the farming practices of UVDs, particularly pumping and diverting the entire stream flow, are significant causes of the degradation of the quality of the Gila River.

D. Extent of Water Quality Degradation

Plaintiffs’ experts, compiling data from several sources, estimate that man-made contributions to salt load account for almost half of the increase in salt in the Safford Valley. See Stetson Engineers, Inc., Preliminary Review of the Salt Balance of Safford Valley 8 (November 3, 1994). The Stetson salt balance study estimates the contribution of salts from various artificial and natural causes, expressed as percentages of the total salt inflow. Stetson Engineers concluded that the artificial contributions include: (1) returns to the river of diverted flows, representing 17% of the salt inflow; (2) returns to the river of pumped underground water, representing 25% of the salt inflow; and (3) artesian wells, representing 6% of the salt inflow. Roughly 36% of the salt inflow in the Safford Valley is. attributed to the salt in the flow when it enters the Valley. Plaintiffs assert that some portion of this salt load is also man-made, stemming from diversions and pumping in the Duncan-Virden Valley.

Defendants’ expert witness, Dr. Matlock, also prepared a salt balance study. Having rejected several of UVDs’ working assumptions and arguments, the Court finds it unnecessary to evaluate the parties’ various arguments related to that study.

However, the Court agrees with defendants that the Stetson salt balance may overstate the effect of UVDs’ farming practices in at least two ways. First, the salt balance study relies on data from 1941 regarding the determination that 25% of all irrigation water, both pumped and diverted, returns to the river as surface runoff. The Court agrees with UVDs that this conclusion may inappropriately discount the effect of more modem farming practices, such as laser leveling of cultivated fields, that reduce surface runoff. Second, the Stetson salt balance reflects the underlying assumption that phreatophytes do not increase salinity at all. However, the evidence unequivocally supports the conclusion that phreatophytes do have an effect on salinity, because they consume large quantities of water and leave behind salt. This has an impact on salt concentration in the remaining water, as well as on localized salt concentrations in the soils. In light of the conflicting testimony, the Court is unpersuaded that the effect is either as great as UVDs suggest or as small as plaintiffs suggest.

The Court cannot conclude that artificial contributions of salt account for a full half of the salt load leaving the Safford Valley. It is sufficient to conclude, for purposes of the issue before the Court, that UVDs’ contributions to the salt load are significant enough to warrant injunctive relief, as described further below.

*1454 E. Effects of Water Quality Degradation

The Apaches argue that the degradation of the water reaching the San Carlos Reservation impedes their efforts to cultivate the types of salt-sensitive crops they once grew. UVDs argue that the effects are overstated and that recent experience justifies the conclusion that the Tribe can raise a variety of crops commercially.

The irrigated agriculture of the Apache Tribe prior to the entry of the Decree included a variety of crops, several of which are salt-sensitive and moderately salt-sensitive. Prior to the turn of the century, the Tribe cultivated and harvested wheat, barley, com, and beans. Testimony at trial established that ancestors of Apaches on the reservation today grew sugar cane, com, melon, cantaloupes, peppers, Apache squash, and a native sweet com, in addition to oats, wheat, hay, and alfalfa and various garden vegetables. The Tribe states that it can no longer grow such moderately salt-sensitive crops as alfalfa, but must cultivate, if anything, salt-tolerant crops such as barley, wheat, and cotton.

UVDs argue that there is a lack of evidence that the Tribe grew salt-sensitive crops commercially. First, UVDs point out that there is little evidence that Apaches grew pinto beans, a particularly salt-sensitive crop. The Court agrees that, although the record demonstrates that Apaches grew beans, it is difficult to conclude that there were large harvests specifically of pinto beans. Upper Valleys farmers testify that pinto beans cannot be grown with Gila River water at all, due to a variety of rust whose spores are present in the stream flow. The Court agrees generally with the Tribe, however, that the historical record amply demonstrates harvests of a variety of moderately salt-sensitive crops, including, but not limited to, com, pumpkins, melons, and cabbage. Equally important, the historical record documents the Tribe’s attempts, often defeated by matters beyond its control (and unrelated to water quality), to cultivate and harvest such crops.

The Tribe and the government presented persuasive evidence that the salinity of the Gila River reaching the San Carlos Reservation today prevents the successful cultivation of such crops as alfalfa, pinto beans, squash, com, melons, and garden vegetables. As the expert testimony at trial indicates, there are two important factors determining crop viability and yield: the quantity of the water and the quality of the water. Because the Tribe, like other parties to the Decree, is limited to an annual water duty of 6 AFA, it is critical that the water applied to the lands be of sufficient quality to permit the successful growth of the crops. Plaintiffs experts, using well-accepted statistical modeling, demonstrated that commercial yields of salt-sensitive and moderately salt-sensitive crops are unlikely unless the quality of the water reaching the reservation improves. 15 Witnesses from both sides agreed, however, that the Tribe can produce full yields of salt-tolerant crops, such as barley, oats, hay, and wheat.

The Court concludes, based on the evidence presented, that plaintiffs have documented the unlikelihood of successful commercial cultivation of salt-sensitive and moderately salt-sensitive crops using Gila River water at its current levels of quality. The Court notes, however, that individual farmers from the upper valleys, such as J.D. Colvin, were optimistic that certain types of crops grown in the upper valleys could, in fact, be cultivated on the reservation. If Mr. Colvin continues his work with the Apache Tribe, it is possible that the dire conclusions of the expert witnesses will be proven incorrect. Until such a demonstration is made, however, the Court must conclude that the effect of the water quality degradation on the Apache Tribe’s efforts to revitalize their agriculture are of a magnitude warranting some form of injunctive relief.

F. Restoring Water Quality at the Reservation

As the above discussion indicates, the Court concludes that injunctive relief is re *1455 quired to restore to the Apache Tribe water of sufficient quality to sustain commercial production of the sorts of crops they grew prior to the entry of the Decree.

UVDs propose extending the Fort Thomas canal (joined with two other canals) to the reservation boundary. An extension of the canal, they argue, would supply the Tribe with water of the same quality as irrigators in the Fort Thomas region receive. The Tribe objects to this proposal for several reasons. First, extending the Fort Thomas canal will not deliver the water to the Tribe’s irrigated lands, but will only get the canal water to the eastern boundary of the reservation. The Tribe is not in a financial position to complete the canal project. Also, UVDs apparently expect the Tribe to join the canal association and pay dues like other users. Second, water from the Fort Thomas canal may not be of the same quality as the Gila River would be, if some amount of the natural flow passed through the upper valleys without being diverted. The Tribe notes that the canal water, by the time it reaches the reservation, may have been applied more than once to cultivated lands and may be heavily supplemented with water pumped from the more saline groundwater. In other words, the canal water may not represent an improvement over the river water. Third, the Tribe will bear significant transportation and seepage- losses from the reservation boundary to the fields. This may defeat the proposal entirely. Fourth, the Fort Thomas Canal has junior rights to some other canals and gets shut down earlier when users are limited to diverting on priority. The Tribe fears that it will forfeit its priority date by being placed on the canal. Fifth, the Tribe asserts its right under the Decree to the natural flow of the river. Even if the doctrine of prior appropriation permits a court to authorize a different point of diversion, an issue which the Court need not and does not reach, the circumstances presented here do not warrant the imposition of a costly and untested approach that may conflict with the Decree. Therefore, the Court will not, under the circumstances presented here, force the Tribe to accept an equivalent amount of water from a source designated by a junior appropriator. 16

The government and the Tribe offer a water quality management proposal drafted by Dr. Gerald T. Orlob. Dr. Orlob’s plan focuses on maintaining EC at 2000 cS/cm at the reservation boundary, a value near or slightly higher than the average salinity found in the Safford Valley. This would theoretically allow the tribe to grow some moderately salt-tolerant crops. 17 Dr. Orlob also proposes attempting to maintain a flow rate of 30 cfs at the reservation boundary by limiting diversions. One objective of his proposal is that the Gila River be maintained as a live stream as far as possible within the natural variability of its flows, and that the stream not be allowed to go dry “as a consequence of diversions.” Finally, Dr. Orlob proposes sealing artesian “hot wells,” discontinuing the use of certain wells of abnormally high salinity, minimizing drainage from irrigated fields to the stream, lining canals that have excessive losses to ground water, and beginning development of both monitoring and modeling of the river and ground water system.

UVDs object to Dr. Orlob’s proposals as unrealistic and possibly devastating to farmers. UVDs argue, for example, that sealing particularly salty wells will ruin certain farmers who rely heavily on well water. The Court agrees with plaintiffs, however, that the river cannot tolerate inflows of extremely salty water. Indeed, the proposal to seal unusually saline wells is somewhat modest.

The Court considers it unnecessary to adopt Dr. Orlob’s proposal in its entirety at this time. Certain aspects of the proposal *1456 would work sweeping changes in the management of the river, some of which may prove unnecessary or unwise. Rather, the Court directs representatives of the affected parties to convene, with the assistance of the Water Commissioner, to discuss alternative methods of altering the management of the river to improve the continuity and quality of flow of the Gila River reaching the San Carlos Reservation. 18 The parties may conclude that they cannot agree on a proposed form of injunction, in which case they may submit proposed forms of injunction for the Court’s review and possible adoption. Any proposed form of injunction must, however, provide for the following:

(1) a goal of reducing EC levels at the reservation boundary to within the range of 2,000 cS/em during the peak irrigation months;

(2) sealing of artesian “hot wells,” such as the artesian well near the Smithville diversion;

(3) limiting diversions in the Salford Valley as necessary to achieve a target flow rate at the reservation boundary. Such a target must be tied in some meaningful way to improving water quality for irrigation purposes. The Court notes that a proposed injunction could, but need not, provide for a gradually increasing target flow rate, with accompanying monitoring of water quality to determine an optimal flow rate for achieving an acceptable level of water quality; and

(4) relatively long-term monitoring and adjustment to ensure not only that the Tribe receives water sufficient for cultivating moderately salt-sensitive crops, but also that the measures taken are no more disruptive in the upper valleys than is reasonably necessary to achieve this goal.

The Court realizes that the Tribe seeks immediate relief on the issue of water quality. The Court concludes, however, that the evidence does not justify issuing an injunction that, for example, requires UVDs to deliver water with EC values of 1,000 cS/cm or less. The Court firmly believes that the parties are in a better position to consider alternative measures for the realistic improvement of water quality in light of the Court’s findings and conclusions set forth above. Further, the Court is optimistic that a measured approach to improving water quality will reveal that the Tribe’s water rights can be satisfied without drastically altering the farming practices in the upper valleys.

Until such time as the parties have’met, conferred, and advised the Court on a proposed course of action, the Court reissues the preliminary injunction entered in September of 1992, Gila Valley Irrigation Dist., 804 F.Supp. at 7, and directs the Water Commissioner to implement the injunction as he had done prior to the vacation of that order by the court of appeals.

II. PRIORITIES AND APPORTIONMENTS

UVDs argue that this Court’s prior determination concerning the Apache Tribe’s 1846 priority right may effectively render their apportionment rights illusory. GRIC argues that the UVDs have no one to blame for this but themselves.

This Court held in 1992 that

the Apache Tribe’s prior right to 6,000 acre-feet of water from the flow of the Gila River is superior in all respects to the rights of the UVDs to the flow of the river. If, at any time, the Apache Tribe asserts its prior right to the natural flow of the Gila River, UVDs may not divert water either on priority or apportionment.

Gila Valley Irrigation Dist., 804 F.Supp. at 7. On appeal, UVDs argued that this holding created the potential for what they call “circular priorities.” See Gila Valley Irrigation Dist., 31 F.3d at 1438. The court of appeals held that the issue was not ripe for adjudication because no call system was yet in place. Id. The appellate court also expressed its optimism that the Rules Corn *1457 mittee would “formulate a system that eliminates the UVDs’ concerns regarding circularity, and preserves their apportionment right, without jeopardizing the Apache Tribe’s priority right.” Id. The Rules Committee did not, however, meet this challenge.

The “circular priorities” effect arises as a result of the juxtaposition of the parties and the relative priority of their calls on the river. Because UVDs’ apportionment call is junior to the Apaches’ priority call, 19 UVDs must allow sufficient water to pass to satisfy the Apaches’ call of up to 12.5 cfs before UVDs may divert apportioned water. Because GRIC’s immemorial priority call 20 is superior to the Apaches’ 1846 priority call, however, GRIC asserts that the Apaches must similarly allow this 12.5 cfs to continue downstream if GRIC issues an immemorial priority call. As envisioned by UVDs, if the Apaches issue a call for 12.5 cfs while GRIC issues a call for 437.5 cfs, then the Apaches will not be permitted to divert any water until the flow exceeds 437.5 cfs, and UVDs will not be permitted to divert any water, even when they have an apportionment available, until the flow exceeds 450 cfs. But if the flow exceeds 480 cfs, UVDs may divert on the ordinary priority schedule, rather than relying on apportionments. Thus, they argue, the “circular priorities” effect renders their apportionments illusory.

This issue is now ripe for adjudication. Under the call system drafted by Gookin Engineers and proposed by GRIC, if UVDs make an apportionment call when both the Apaches and GRIC make priority calls on the river, then GRIC’s call must be satisfied before the Apaches’ call is satisfied, and the Apaches’ call, in turn, must be satisfied before UVDs’ call. This leaves the UVDs unable to divert on apportionment unless the flow exceeds 450 cfs. UVDs’ fears of the “circular priority” effect are therefore no longer speculative. Moreover, GRIC’s proposal conflicts with that of the Water Commissioner. The Water Commissioner suggests a different procedure: UVDs would be permitted to divert on an available apportionment up to 437.5 cfs, but would be required to let pass any amount over 437.5 cfs to satisfy a call made by the Apaches. The Court understands that under the Commissioner’s proposal the Apaches would be able to divert on their 1846 call before GRIC could divert on their immemorial priority call when UVDs divert on apportionment. But as between the Apaches’ 1846 call and the UVDs’ apportionments, the Commissioner assigns the higher priority to the apportionment. Given the need to implement a call system as soon as practicable, the Court concludes that the “circular priorities” issue must be resolved.

GRIC argues that the “circular priorities” effect is a predictable and unobjectionable consequence of both the language of the Decree and the process by which it was drafted. First, GRIC argues that because apportionments are exceptions to the priority schedule, they must be strictly applied in a manner that will not otherwise disrupt the priorities among other parties. More importantly, perhaps, GRIC argues that UVDs were the agents of their own discontent. GRIC relies on letters and memos written by the parties negotiating the Decree in the early 1930s. The negotiating parties devised a “substituted storage” plan under which UVDs would be permitted to divert water in disregard of the United States’ prior rights, including those held on behalf of the San Carlos Apache Tribe, provided UVDs paid a fee of $10.00 per acre for forty thousand acres of land in the upper valleys. 21 UVDs, however, *1458 ultimately refused to pay this fee. For example, an April 19, 1933 joint memorandum signed by several of the negotiators 22 stated that “[i]n recent conferences, it has appeared that the Upper Valleys do not want, and, in fact, refuse to make any payment on account of the substituted storage feature of the decree, other than agreeing to the settlement of the suit on the lines set forth in the decree.” The proposed Decree language was subsequently modified to provide for apportionments to be made in disregard of the prior rights of the United States below the reservoir, thereby excluding the Apache Tribe’s 1846 right from the apportionments clause. See Article VIII(2). The Decree also provided, however, that UVDs would be able to purchase and transfer the Apache Tribe’s right to the San Carlos Irrigation Project for $62,500, if certain conditions were met. Article VI(2). UVDs did not make this purchase. GRIC argues that the UVDs could have obtained the transfer of the Apache Tribe’s right to SCIP, but chose not to, and that they must therefore bear the consequences of their decision.

The Court rejects this reasoning. First, as a practical matter, GRIC’s approach renders the apportionments clause illusory. As noted above, if GRIC is permitted to exercise its immemorial priority right by virtue of the precedence the Apaches’ 1846 priority has over UVDs’ apportionment calls, then UVDs will be able to divert on apportionment only when flow in the river exceeds 450 cfs, regardless of the amount of water stored in the reservoir and available for use downstream. As UVDs demonstrated at trial, based on data from a twelve-year period, this approach would prevent them from irrigating during the bulk of the season. Indeed, the median flows between 1980 and 1992 at AshurstHayden Dam never exceeded 450 cfs between mid-May and early December. Further, as is also noted above, when the flow in the river reaches 480 cfs, UVDs’ priority rights are triggered, largely obviating the need for apportionment calls. In effect, GRIC’s interpretation of the Decree eliminates the benefit of the apportionments clause by superimposing upon it the doctrine of prior appropriation. This result is at odds bo