JUSTICE ERICKSON

delivered the opinion of the Court.

Plaintiff-appellant Joe Berens appeals from a district court decision which affirmed the Colorado Ground Water Commission’s (Commission) rejection of his application to construct a well. The district court held that there was no unappropriated ground water available in the three-mile circle surrounding Berens’ proposed well site and that further appropriations would unduly deplete the aquifer to the detriment of senior appropriators. We remand this case to the trial court for a further hearing and findings of fact on the issues set forth in this opinion.

On March 27, 1974, Jack R. Bond filed an application with the Commission for permission to construct a well in Kit Carson County. The application requested permission to pump 400 acre-feet per year (afy) from the Ogallala formation for the irrigation of 480 acres of land located in the Northern High Plains Designated Ground Water Basin. Subsequent to that filing, Joe Berens succeeded to Bond’s interest and became a replacement .applicant for the proposed well (the Berens well.) 1

*172 On May 31, 1974, the deputy state engineer notified Bond that his application had been rejected based on the results of a preliminary evaluation conducted by the Commission staff. The Commission’s evaluation, which was conducted prior to the issuance of our opinion in Thompson v. Colorado Ground Water Commission, 194 Colo. 489, 575 P.2d 372 (1978), consisted of a comparison of the total amount of ground water available for appropriation with the sum of those amounts claimed by senior appropriators in their applications and conditional permits. The Commission’s evaluator determined that there were 5,500 afy of ground water available for appropriation in the three-mile circle surrounding the Berens well (the Berens circle), and that the total claimed appropriation equalled 7,593 afy.

Thereafter, Bond and replacement applicant Berens requested a hearing before the Commission to contest the denial of their application. See section 37-90-114, C.R.S. 1973. At the conclusion of that hearing, the Commission again denied the application and an appeal was taken to the district court. See section 37-90-115, C.R.S. 1973.

A trial de novo was held in the district court on April 4, 1979. After hearing conflicting testimony from expert witnesses, the trial court affirmed the Commission’s rejection of the well application, finding “that the water requirements for the acres actually under irrigation exceed the water available in the aquifer if the forty percent depletion in twenty-five years criterion is to be accomplished.” The court stated that there were approximately 3,100 acres presently under irrigation within the Berens circle.

The district court did not make specific findings on the amount of water available for appropriation in the Berens circle and the amount of water subject to claims by prior appropriators. Because these factors play a central role in the Commission’s 40% depletion in 25 years formula, we are not in a position to review the trial court’s determination that the Ber-ens circle was overappropriated. Accordingly, we remand this case to the trial court to make further findings on the issues which are set forth in this opinion

II.

In Fundingsland v. Colorado Ground Water Commission, 171 Colo. 487, 468 P.2d 835 (1970), this Court first recognized and approved the Commission’s three-mile, 40% depletion in 25 years formula: “The three mile test was developed for use in the Northern High Plains. It is partly based on policy and partly based on fact and theory. Using that test, a circle with a three mile radius is drawn around the proposed well *173 site. A rate of pumping is determined which would result in a 40% depletion of the available ground water in that area over a period of 25 years. If that rate of pumping is being exceeded by the existing wells within the circle, then the application for a permit to drill a new well may be denied. The formula again received our approval in Thompson v. Colorado Ground Water Commission, 194 Colo. 489, 575 P.2d 372 (1978). In Thompson, however, we disapproved the Commission’s practice of computing the ’’quantity of existing claims“ based on the amounts of ground water claimed on the face of conditional permits. We held in Thompson that the Commission must determine the actual amount of water being put to beneficial use by the holders of conditional permits. That amount, when added to amounts claimed in final permits, could then be used to determine the quantity of existing claims senior to the applicants.

Peterson v. Ground Water Commission, 195 Colo. 508, 579 P.2d 629 (1978), addressed a further question arising from our rejection of the Commission’s practice of evaluating well permit applications under the assumption that all conditional rights to appropriate were being fully exercised. In Peterson, we declared:

“Holders of conditional permits may not claim the maximum amount permitted under their permits and yet place only a portion of the claimed water to beneficial use. . . . Water claimed under conditional permits which have expired is validly appropriated only to the extent of actual beneficial use prior to the expiration of the conditional permit, regardless of the fact a final permit based upon the extent of beneficial use has not been issued.”

Our reasoning was founded upon the statutory scheme established in the Colorado Ground Water Management Act, section 37-90-101, C.R.S. 1973, et seq., which provides conditional appropriators with a one-year period in which to put their appropriation to beneficial use. See section 37-90-108(3), C.R.S. 1973. 2 In our view, the Act does not contemplate that appropriators may acquire a vested right in water put to beneficial use after their conditional permits have expired.

III.

Although we are remanding this case to the trial court, further discussion of the issues to be addressed is necessary to facilitate the proceedings below.

A. Quantity of Ground Water Available for Appropriation.

The trial court must initially determine the amount of ground water that is available for appropriation under the Commission’s 40% depletion in 25 years formula. Prior to our decision in Thompson, the Commission *174 employed what may be called its “standard” three-mile formula to determine the amount of ground water available for appropriation. Under the Commission’s formula, such factors as the saturated thickness of the aquifer, its specific yield, precipitation recharge, and the irrigated return accruing to the aquifer from surface irrigation were considered. In each evaluation, the Commission employed basin-wide values for all of the variables used in its formula except saturated thickness, which was determined through a series of tests conducted within the three-mile circle under examination. As a result, the Commission’s working formula computed ground water available for appropriation as a function of saturated thickness.

According to testimony by the Commission’s expert witness, Leonard Mercer, the standard three-mile formula did not, in practice, result in a 40% depletion rate in 25 years. Mercer stated that, because a small shift in a proposed well site could eliminate from consideration those senior wells lying on the perimeter of a fully-appropriated circle, and cause the acceptance of the proposed well, the circle could, as a result, be overappro-priated. 3 In support of his position, Mercer cited Commission studies suggesting that a number of fully-appropriated townships were actually over-appropriated by about 40%.

Although Commission experts were aware that the standard formula theoretically resulted in an overappropriation of ground water, they concluded that any actual overappropriation would be negated by the Commission’s policy of using claimed appropriation figures (as stated on conditional permits) in their calculation of actual appropriation. Since these figures grossly overstated the amount of water actually being appropriated, their use resulted in an underappropriation of the three-mile circle which compensated for the overappropriation inherent in the Commission’s standard three-mile formula.

As set forth in part two of this opinion, we disapproved the Commission’s use of claimed appropriation figures in Thompson v. Colorado Ground Water Commission, supra. As a result, the Commission has recently modified its standard three-mile formula to take into account the *175 percentage of overappropriation thought to be inherent in that formula. 4 Based on the Commission studies suggesting that a 40% overappropriation resulted from the use of the Commission’s standard formula, the Commission modified the formula to achieve a 40% reduction in its estimate of the amount of available ground water for each three-mile circle. 5 In the Berens circle, for example, the Commission revised its estimate of available ground water downward from 5,500 afy to 3,302 afy. 6

On remand, the trial court is directed to make specific findings concerning the Commission’s modified formula, including a determination as to whether the estimate of 40% overappropriation is reasonable. The court may conduct further hearings on this issue if necessary and should also consider the reasonableness of any other modifications which the Commission made in the formula after the announcement of Thompson. As we have pointed out in note 4, the applicant’s estoppel argument is rejected as a matter of law and need not be considered on remand.

B. Quantity of Existing Claims.

Upon determining the amount of ground water available for appropriation in the Berens circle, the trial court must then make a specific finding as to the quantity of existing claims senior to Berens’ application. Inherent in this determination are two issues relating to the three-mile circle: How many acres are being irrigated in the three-mile circle and how much ground water is being applied to each acre under irrigation? Only after specific findings have been made on each issue may the trial court reach a conclusion as to the quantity of existing claims senior to Berens’.

With respect to the first issue, the trial court found that there were approximately 3,100 acres under irrigation at the time of trial. Although Berens does not contest that determination, he argues that the trial court should have determined how much acreage was under irrigation at the time all conditional permits senior to Berens’ expired. Charles Hallenbeck, Berens’ expert witness, testified that only 1,284 acres were under irrigation in 1975, when, according to Berens, most or all of the permits senior to Berens’ expired:

In Peterson v. Ground Water Commission, supra, we held that water claimed under a conditional permit which has expired is valid only to the extent that the water is appropriated and put to beneficial use prior *176 to the expiration of the conditional permit. Accordingly, on remand the trial court is directed to determine when and if the conditional permits senior to Berens’ application expired. Once it has been determined that some or all of the conditional permits senior to Berens’ have expired, the trial court must make an additional finding as to the number of acres under irrigation at the time those permits expired. Of course, where conditional permits have not expired as of the date of trial, the trial court should assume that the full conditional appropriation will be put to beneficial use.

The second issue to be resolved in determining quantity of existing claims concerns the amount of ground water actually applied to each acre under irrigation. Although we have not previously required that the Commission adopt any particular procedure for making this calculation, the clear import of our holding in Thompson was that the Commission should take all reasonable steps to determine the actual amount of water applied per acre. Presently the Commission allows each appropriator to claim an average amount of water (30“), which is derived from a composite of such factors as average rainfall, consumptive use, and the efficiency of the irrigation method chosen. The question to be addressed by the trial court on remand is whether the Commission’s policy of allowing each appropriator to claim an average amount of ground water squares with the doctrine in Thompson.

At trial and on appeal, the Commission argues that its averaging policy is the only feasible means available for determining the amount of ground water put to beneficial use by appropriators. Because the actual amount of ground water subject to appropriation may be expected to vary each year, depending on the crops being planted and the amount of rainfall that occurs, the Commission asserts that any attempt to measure the actual amount of ground water put to beneficial use during the one year life span of the conditional permit would result in an arbitrary distribution of ground water. During years of low rainfall, for example, conditional appropriators would put larger amounts of ground water to beneficial use and would receive a vested right in their abnormally high appropriation. Conversely, during a year marked by high rainfall, conditional appropriators would put smaller amounts of ground water to beneficial use and would lose their right to appropriate larger amounts in later years when there is less rainfall. It is the Commission’s position that neither result is consistent with the legislative intent underlying the Colorado Ground Water Management Act, section 37-90-101, C.R.S. 1973, et seq.

In considering the propriety of the Commission’s averaging approach, the trial court should carefully consider the evidence pertaining to the Commission’s choice of 30” as an average figure for appropriation in the Northern High Plains Basin. In Thompson, we expressed doubts about this figure and, although we recognize the difference between application *177 and consumptive use, we are concerned that the Commission’s figures may be overstated. Berens’ expert witness presented persuasive evidence to this effect.

Accordingly, we remand this case to the trial court for further hearings and for specific findings on the issues set forth in the course of this opinion.

In a letter addressed to Joe Berens, filed with the State Engineer on October 21, 1974, Bond relinquished his interest in the well application in favor of Berens. A postscript to that letter also contained a promise to sell Berens one acre of land at the proposed well site. Subsequent to making that promise, Bond sold the portion of his land which had been designated as the well site to Bob Strieker, who testified that he had made no arrangement to sell land to Berens. In the district court, Berens testified and the court found that it was his intention to ask for leave of the Commission to move the well if the application was granted. Although we recognize that there is a question under section 37-90-107, C.R.S. 1973, as to whether the well permit accurately reflects the ownership of the proposed well site, that issue was not raised on appeal and we do not address it here.

The one-year limitation imposed by section 37-90-108 has recently been expanded to three years. See section 37-90-108, C.R.S. 1973 (1979 Supp.).

The following example, based on the testimony of Leonard Mercer during Commission hearings, may help to clarify the Commission’s argument on this issue. Assume that applicant A was denied a well permit on the grounds that a three-mile circle around his proposed site is fully appropriated. Assume also that there are several wells near the east and west perimeters of A’s circle. If applicant B applies for a permit at a location 3/4 of a mile to the east of applicant A’s location, the resultant shift eastward in the circle would eliminate all of the western wells from consideration, making B’s circle underappropriated. Applicant B’s permit would be granted while A’s would still be denied. In addition, applicant C could also get a permit based on a well site that was 3/4 of a mile to the west of applicant A’s site. With the addition of wells for B and C (and potentially others), the original circle would become greatly overappropriated.

Berens argues that the Commission should be estopped from substituting its new formula for the standard formula upon which it relied to deny Berens’ application. In Colorado Ground Water Commission v. Dreiling, 198 Colo. 560, 606 P.2d 836 (1979), we recognized that equitable estoppel may be applied against government agencies to prevent manifest injustice. There, however, we held that the applicant had not reasonably relied on the Commission’s actions. In this case we similarly find no reasonable reliance and therefore reject Berens’ claim.

According to evidence in the record, the Commission simply divided its standard formula by 140% in order to derive its new formula.

The Commission has also modified several of the figures used in its three-mile formula to reflect more current scientific information relating to the water-bearing characteristics of aquifers in the Northern High Plains Basin.