AN ANONYMOUS VIEWER RESPONSE FORM AND OTHER WEBSITE MATTER ARE IN THE FOOTER. A RESPONSE IS APPRECIATED.

Groundwater Levels and Groundwater Chemistry Gallatin Valley Montana 1977

Prepared for Blue Ribbons of the Big Sky Country Areawide Planning Organization, Bozeman, Montana

by Earth Science Services, Inc.,

Darrel E. Dunn, Principal Investigator

January 9, 1978

[This is an html version of the report that was originally published in paper copy. Transcribed by Darrel Dunn, Ph.D., Hydrogeologist, Colorado Springs, Colorado.]

INTRODUCTION

The U. S. Geological Survey completed an investigation of the water resources of the Gallatin valley in 1954, and they published the results of the investigation in Water Supply Paper 1482 titled Geology and Ground Water Resources of the Gallatin Valley, Montana (Hackett et al, 1960, hereafter called WSP 1482). This publication is the major source of information on groundwater conditions in the Gallatin Valley. However, present groundwater conditions are not necessarily the same as they were twenty-three years ago, because the ground water system is influenced by factors that may change. Consequently, it seemed prudent to check the current validity of the groundwater system described in WSP 1482, and also to check for any indication of significant changes in the chemical quality of the groundwater.

This report describes an investigation that was designed to serve these purposes by checking the general validity of the maps and findings of WSP 1482 with respect to present groundwater conditions in the valley. The present investigation was not, however, designed to discover or study all of the local ground water changes that might have occurred.

In the following report, the methods of the investigation will be described first and then the findings will be presented.

METHOD

Water Level Measurements

The U. S. Geological Survey measured the water levels in 234 wells in the Gallatin Valley during 1951-53 (WSP 1482, Table 36). They made periodic measurements on 124 of these wells (WSP 1482, Table 35). In the present investigation, we tried to find as many of these previously measured wells as practical and measure their current water levels for comparison with water levels in the 1951-53 period. Seventy-six of the wells were found and measured, and an additional 18 wells were measured as alternates to wells that could not be measured and to supplement the data of WSP 1482. Thirty-two wells and 7 alternates to the wells that the Geological Survey measured periodically (Table 34, WSP 1482) were measured twice; the first measurements were made in early July, and the second in late August and early September, 1977. We made nearly all of the measurements by chalked steel line, but a few were made with an electric probe. We used U. S. Geological Survey measuring points on the wells where they still existed. Table A-1 contains the water level measurements. The measuring point elevations given in Table A-1 are from WSP 1482 (Table 36) or, where no elevations are given in that table, they were taken from the U. S. Geological Survey 15-minute topographic quadrangle maps for the area.

Well Number System

The well number system used in this report is the system used in WSP 1482. The first letter (capital) of the number indicates the quadrant of the principal meridian and base-line system in which the well is located. The letter "A" indicates townships north and ranges east; and "D" indicates townships south and ranges east. None of the wells fall in quadrants "B" and "C." The first numeral of the number denotes township; the second, the range; and the third, the section. Lowercase letters following the section number indicate, respectively, the quarter section, the quarter-quarter section, and so on. These subdivisions of the sections are designated a, b, c, d and are assigned in counterclockwise direction beginning in the northeast quarter. Hence, a well located in NE 1/4 of SE 1/4 , section 16, T1N, R4E is numbered A1-4-16da. This system is illustrated on page 10 of WSP 1482

Water Sample Collection

The U. S. Geological Survey performed chemical analyses on groundwater from 34 wells and 3 springs in the Gallatin Valley (WSP 1482, Table 27). In the present investigation, we sampled 17 of these wells, and 15 additional wells were sampled as alternates for wells that were not available for sampling or could not be sampled with a reasonable amount of effort. Most of the alternates were close to the locations of the wells originally sampled by the U.S. Geological Survey. In addition to the wells, all three springs were sampled in the present investigation.

The procedure for collecting samples from wells was as follows:

Prior to collecting the sample, each well was pumped for 1 hour or until a quantity of water equal to approximately twice the volume of the well and the pressure tank was produced, whichever occurred first. This amount of pumping was reduced in some cases where the well user had pumped the well just before sampling. In 3 cases, the well was not pumped prior to sampling, and we collected a water sample by using a small, hand-operated pump. One domestic well was pumped only enough to obtain a sample; this was in accordance with the owner's request.
The pH and electrical conductivity of the water was measured at the well sampling site.
Samples were collected in clean polyethylene bottles. We rinsed the bottles with the well water, filled them to the top, and capped them tightly.
The samples were not filtered or acidified at the well. They were placed in a cooler, and delivered to a laboratory at Montana State University within 8 hours of the collection time. We followed the same procedure for spring water, except that no pumping was needed.

The water samples were analyzed at the Montana State University Chemistry Station Analytical Laboratory in accordance with U. S. Environmental Protection Agency specifications (Environmental Protection Agency, 1976), and Table A-2 contains the results.

WELL WATER LEVELS

Significance

Water levels in wells are a measure of the hydraulic head in the aquifer, and groundwater moves in the direction of decreasing head in accordance with established principles of groundwater movement. As an approximation, the groundwater may be expected to be moving in a direction that is perpendicular to the contour lines on Plate 1. Consequently, Plate 1 can be used to estimate the general direction of movement of groundwater and any mobile impurities that it might contain.

Changes

Inspection of Plate 1 shows that no water level changes have occurred since 1953 that would appreciably change the general direction of groundwater flow. Indeed, the differences shown are probably within the normal groundwater fluctuations that result from ordinary variation in recharge from place to place, season to season, and year to year. The largest difference in water levels are in the vicinity of Belgrade where several wells have lower water levels than they did at corresponding times of the year in 1951-53 (see Plate 1 and Plate 3). These lower water levels are probably a result of the abnormally low supply of irrigation water available in 1977 that resulted from low snowpack accumulation in the mountains the previous winter.

A hypothesis that explains the relatively large water level difference near Belgrade is that the area around Belgrade is subject to larger recharge related fluctuations than most of the rest of the Gallatin Valley. Such fluctuations may be seen on Plate 7 of WSP 1482, where water level rises as great as 40 feet are shown to have occurred during the irrigation season. The area of these fluctuations coincides with the area underlain by thick highly permeable Quaternary alluvium. The natural groundwater levels in this area are low. This may be seen by inspecting the April contours on Plate 5 of WSP 1482. These contours show water levels that are controlled by the levels of the Gallatin River and the East Gallatin River, with some of the flow of the Gallatin River being diverted through the gravels in the Belgrade area. However, during the irrigation season, downward percolating irrigation water causes the well water levels to rise to elevations considerably above their natural levels; and the subsurface flow from the Gallatin River is reduced. The resulting groundwater regimen seems to be one of large annual water level fluctuations that are controlled by the amount of irrigation water applied. Information in the Gallatin County Water Resources Survey (Montana State Engineer's Office, 1953) suggests that the amount of irrigation water applied in the Belgrade area will decrease considerably in years of water scarcity because much of the acreage is supplied by junior water rights. For example, a large part of the area is shown as being supplied (in 1953) by the Spain-Ferris Ditch Company whose oldest rights are from 1886 appropriations. These are regarded as high-water rights, which entitle the irrigators to water only during high flow periods of the West Gallatin River. Consequently, since the irrigation water supply was low in 1977, it is not surprising that water levels in the Belgrade area were low.

Other explanations for the low water levels near Belgrade were considered. One possibility that was investigated was that the data was misleading because of the use of alternate wells and wells that were being pumped. However, a well by well analysis showed that this was not the case, and one well (D1-5-9cd) is a U. S. Geological Survey observation well that is not used at all and is not located close to any other well that might affect it. Another possibility considered was that the pumpage of municipal and industrial wells at Belgrade might have caused the lower water levels. However, calculations using an equation for drawdown (decrease in water levels due to pumping) in water table aquifers (Walton, 1970, p. 222) strongly suggest that such pumpage would not cause large declines. These calculations are in the appendix of this report. Furthermore, there are no obvious continuous declines in water levels in two U. S. Geological Survey observation wells that are located near Belgrade in the same aquifer as the Belgrade municipal wells. One of these wells is A1-4-25dc, which is about 1 1/4 mile north of Belgrade. Tables of water levels in these wells are included in the appendix.

Most of the remainder of the Gallatin Valley also had lower water levels in 1977 than 1951-1953, but the difference was less than in the Belgrade area. Inspection of the precipitation and streamflow graphs in Figures A-1 and A-2 will show that a valley-wide decline in water levels is not caused by an extended drouth. Instead, these relatively low water levels are probably due to the reduced amount of irrigation water available in 1977. However, some wells on the Bozeman Fan in the south half of T2S, R5E show water levels that are slightly higher than in 1951-1953. These wells are in an area that is shown in the Gallatin County Water Resources Survey (Montana State Engineer's Office, 1953) as being served by the Middle Creek Ditch Company, which is described as having mostly senior water rights. The high water levels are probably due to senior water rights plus chance measurement soon after a heavy application of irrigation water.

One anomalous well is D2-4-9bc. This well was equipped with a water level recorder from September 1951 to December 1952, and daily records show that the depth to water was never less than 22.57 feet; but on September 15, 1977, the depth to the water level was only 17.28 feet. The cause of this rise is not known. However, a rural home was built near the well subsequent to 1952; and an irrigated lawn is present now. Another possible explanation is that the well was plugged back after 1952,and that the well is in a ground water recharge area where hydraulic head decreases with depth.

The water level measured in D1-4-14bb on July 14, 1977 is also anomalously high when compared with the water level measured in the same well on July 11, 1951. The explanation for this large positive water level difference may be that irrigation began later in 1951 than in 1977. In 1951 the water level was at 19.31 feet below the measuring point, but it rose 10.79 feet by August 2. This rapid rise suggests that irrigation began in the last half of July. Consequently, the data suggests that the July water level was higher in 1977 than in 1951 because irrigation had begun in the vicinity of the well in 1977 but it had not begun yet in 1951. Some other positive July numbers on Plate 3 may have a similar origin.

GROUNDWATER QUALITY

Total Dissolved Solids

Data on the quality of the groundwater in the Gallatin Valley is contained in Table 27 of WSP 1482, which is reproduced in the appendix of the present report (Table A-3). This table contains values for concentrations of the major ions commonly found in groundwater plus potassium, silica, iron, fluoride, nitrate, boron and pH. The latter constituents are also common in groundwater but are generally present in minor amounts. The concentrations of all of these elements may be expected to vary due to natural causes. The causes are complex, and the reader is referred to Hem (1970) and Davis and DeWeist (1966) for reviews of the factors affecting the chemical composition of groundwater. The variation in concentrations in the Gallatin Valley includes (1) changes in the chemical composition of the groundwater from place to place in the valley, (2) changes with depth below the ground surface, and (3) changes with time. Such changes may be seen in the chemical analyses published in WSP 1482, and they are depicted in Figures A-3, A-4, A-5, and A-6 of the present report. As part of the present investigation, the chemical data gathered by the U. S. Geological Survey in 1951-53 and published in WSP 1482 was compared to chemical analyses for groundwater samples gathered in 1977 to see whether any significant changes could be detected. Most of the 1951-53 samples were collected in late August and September; the 1977 samples were collected in September. This correspondence in time should reduce the effects of seasonal water chemistry variations. The comparison of total dissolved solids values is shown graphically in Figure 1, and comparisons for the individual constituents are shown on graphs in the appendix. The comparisons are divided into three classes:

resampled wells, which are cases where the same well sampled by the U. S. Geological Survey in 1951-53 was resampled in 1977;
nearby alternate wells, which are cases where an alternate well was sampled at the homesite where the well sampled in 1951-53 was located; and
distant alternate wells, which are some distance from the well sampled in 1951-53.

Comparison of total dissolved solids concentrations, Gallatin Valley, Montana

The distant alternate wells are all within 1 1/2 miles of the original well sampled, and most are less than 3/4 mile away. The concentrations in this report are expressed in parts per million (ppm) so that they will be consistent with the units used in WSP 1482. However, the chemical analyses made in the present study were reported in milligrams per liter (mg/l) which is very close to being numerically equivalent to ppm (Hem, 1970, p. 80).

In Figure 1, the points above the diagonal correspond to instances of increases in total dissolved solids concentrations measured, and the points below the diagonal correspond to decreases in concentrations. The fact that increases are more numerous than decreases suggests that there has been a general tendency for dissolved solids concentrations to rise within the last 24 years. This tendency is suggested even more strongly by the results from the resampled wells. These results are easier to interpret, because the recent sample came from the same well sampled in 1951-53. The average value for total dissolved solids in these wells was 229 ppm in 1951-53 and 259 ppm in 1977. The average increase was 30 ppm. The t-distribution (Dixon and Massey, 1957, p. 115) was used to investigate the probability that this average increase is real. The alternate possibility is that the increase is due to sampling error and natural temporal variation in a system that has not really changed. The result of the application of the t-distribution was that the probability of an actual increase is greater than 90 percent. This increase, if real, is a general valley-wide tendency, because the locations where increases are recorded are scattered over the entire area.

Two of the resampled wells have been excluded from this analysis because they were not pumped before they were sampled, and they contained stagnant water that may have lost bicarbonate as a result of rusting of the steel casing. The proposed mechanism is that ferrous ions from the casing pass into the stagnant well water. These ions then react with bicarbonate ions and dissolved oxygen to produce a ferric compound. A convenient representation is (Davis and DeWiest, 1966, p. 101):

2Fe2+ + 4HCO3- +H2O +1/2O2 = 2Fe(OH)3 + 4CO2

The true reaction may be more complex, but the reduction of bicarbonate is in agreement with the chemical composition of the stagnant water taken from well number A1-4-19cb, in which the iron concentration is high and the bicarbonate concentration is low. Reducing the bicarbonate concentration would tend to reduce the pH in accordance with

CO2 + H2O = H+ + HCO3-

(Davis and DeWiest, 1966, p. 102), and the pH is, indeed, anomalously low. It was measured at 6.8 in the field and 5.5 in the laboratory. The drop in pH is in accordance with additional oxidation of ferrous iron during transport of the sample. The reduction in concentration of calcium, magnesium, boron, nitrate, potassium, and sodium in the well is not explained by these reactions. Dilution by rainwater seems unlikely because the well is capped. Perhaps these ions are being used by microscopic organisms present in the stagnant water; adsorption and organic complexes and colloidal material might also be involved.

In attempting to explain the probable valley-wide increase in dissolved solids, the following hypotheses were considered:

It is not a real increase, and is only a result of sampling from a variable system.
It is not a real increase, and is caused by differences in sampling and analytical procedures.
It is caused by increase in the number of residences disposing of wastes through septic tanks.
It is caused by the 1977 water deficiency.
It is caused by irrigation water leaching minerals from the subsurface materials.
It is caused by increase in dry land farming areas.
It is caused by changes in agricultural practices.
It is a result of a combination of some of the above causes

Before discussing these hypotheses, it should be mentioned that the groundwater in the valley appears to be saturated with respect to calcite (calcium carbonate), and this condition tends to inhibit increases in dissolved solids, because any increase in dissolved calcium or bicarbonate will tend to cause calcite to be precipitated. Consequently, if the soluble minerals available for leaching are gypsum (CaSO4 . H2O), hydrated species of magnesium carbonate, or other minerals containing either calcium or carbonate, then solution of these will tend to be partly balanced by precipitation of calcite. The condition of saturation with respect to calcite is indicated by calculations of equilibrium pH in relation to calcium and bicarbonate activities in solutions in contact with calcite. The procedure followed is described by Hem (p. 133-134) and involves the use of the Debye-Huckel equation to obtain activities for the calcium and bicarbonate ions. The calculations were made for three samples collected in September, 1977 from (1) D2-5-14ac on the Bozeman Fan, (2) D3-4-3ab1 on the Valley Floor, and (3) D2-3-11aa in the Camp Creek Hills. The results are summarized in Table 1. The pH measured in the field is probably less accurate than the pH measured in the laboratory, but it has the advantage of being measured in water immediately after it was pumped from the well, thus reducing the chance that pH will increase as a result of loss of CO2 from the water. All of the measured pH values are higher than the equilibrium pH except the field pH for D3-4-3ab1, and it is possible that this field measurement is slightly low due to difficulty in precisely calibrating the field instrument. The relatively high measured pH values suggest saturation or supersaturation. However, the presence of magnesium may complicate matters with regard to the solubility of calcite (Hem, 1970, p. 142).

With regard to hypothesis number 1, a one-tailed t-test (Dixon and Massey, 1957, p. 118) was used to test the hypothesis that the differences between dissolved solids in the resampled wells came from a population in which the true mean value is less than or equal to zero. The level of significance used was 0.10 and the hypothesis was rejected. This may be interpreted to mean that when the variation in the differences is taken into consideration, there is less than one chance in ten that there is not an actual increase in the true average value for total dissolved solids in the Gallatin Valley area. Thus, an element of uncertainty exists, but it is fairly likely that a real increase has occurred. The use of the t-test seems reasonable, because the cumulative frequency distribution of the differences approximates a straight line on normal probability graph paper.

With regard to hypothesis number 2, the sampling procedures were different. One important difference is that the wells were probably not pumped as much before sampling in 1951-53, because most of the wells were still equipped with hand pumps. In 1977 the wells were pumped as much as 30 minutes at rates that exceeded 5 gallons per minute in some cases. This difference in pumping time appears to have caused more iron to be present in the 1951-53 samples than in the 1977 ones. The second important difference is that the 1951-53 samples were probably collected in glass bottles, whereas the 1977 samples were collected in plastic ones. This difference is thought to be the reason that 1951-53 samples are much higher in silica than the 1977 samples. However, these differences do not directly affect the dissolved solids concentrations given above, because those concentrations were calculated without including iron and silica. Excluding these constituents is unlikely to significantly affect the values calculated because the 1977 concentrations of both were very low, except that iron was high in the wells that were not pumped. However, these unpumped wells were not included in the analysis of the dissolved solids or in the analyses of other constituents that will be described later. With regard to the values for constituents other than iron and silica, no reason for suspecting any bias resulting from sampling or chemical analysis was found. The recent analyses for some of the ions are accurate and reported with more significant digits than the 1951-53 analyses, but an increase in accuracy should not bias the results. The chance that hypothesis number 2 explains the difference seems small, but it cannot be rejected with complete confidence.

With regard to hypothesis number 3, if the septic systems were causing an increase in dissolved solids, the wells in the growth areas, where additional septic tanks have been installed since 1954, should be associated with increases; and wells in areas that have not been developed should be relatively unchanged. This relationship does not seem to be indicated by the data. Well number D2-4-11dc is in a developed area along Highway 289 about a half mile west of the junction with Highway 191 (4-corners). The total dissolved solids in this well were 225 ppm in 1953 and remained at exactly the same value when resampled in 1977. Well number D1-5-22cd is in a developed area near Interstate 90 about 3 miles northwest of Bozeman; the owner reported that this well has been deepened from 30.1 feet at the time it was sampled in 1953 to about 75 feet, but the dissolved solids show in increase of only 8 ppm. Well number D1-4-13bb is located in a developed area along Highway 291 (Jack Rabbit Road) just south of Belgrade. This well is not the same one originally sampled, and it is reported to be 120 feet deep whereas the original well was only 65 feet deep. The dissolved solids in the original well were 226 ppm compared to 407 ppm in the recently sampled alternate well. Another alternate well is located at D3-5-3da just north of a growth area near Hyalite Creek about 5 miles south of Bozeman. The original well, which has been destroyed, contained water with 131 ppm dissolved solids. The recently sampled well is a new well at a recently built home located about 100 yards from the old well. The new well contains water with 127 ppm dissolved solids. These four wells are all located so that groundwater from the developed areas would be expected to flow toward them. To summarize the findings for wells in or adjacent to developed areas, the one resampled well showed no change; the deepened well showed a slight increase; one of the nearby alternate wells showed a large increase; and the other alternate showed a slight decrease. Many of the wells that are not located near developed areas show increases in dissolved solids, and the greatest increase in dissolved solids in a resampled well was found in number D1-3-16aa, which is in an undeveloped area west of Amsterdam. Water from this well increased from 290 ppm in 1951 to 441 ppm in 1977. It has not been deepened. It seems very unlikely that septic tanks have significantly contributed to any general rise in dissolved solids.

With regard to hypothesis number 4, natural fluctuations in dissolved solids might be caused by variations related to the amount of time the groundwater has been in contact with soluble and unstable minerals. Such minerals might include gypsum, hydrated species of magnesium carbonate, ferromagnesian minerals, sulfides, and feldspar. Reaction rates for some of these minerals are slow so that dissolved solids concentrations in water percolating downward toward the water table are likely to be highest when water movement is slow and more time is available to allow the dissolved solids to build up. The dissolved solids concentration of water entering the groundwater system is then related to the rate of percolation of water to the water table and to the amount of subsequent dilution. The dilution can be caused by slugs of water moving rapidly downward to the water table as a result of infiltration of large quantities of water in short periods. The water could result from spring snowmelt and rainfall or large applications of irrigation water. Dilution could also occur as a result of mixing with water that has infiltrated from streams and is moving toward discharge areas in the groundwater system. The interrelationships of these factors as they affect dissolved solids is complex and not predictable. However, dry weather, irrigation water shortages, and low spring streamflow might tend to produce a temporary rise in the dissolved solids concentration. Such water deficiencies were experienced in 1977. These recent deficiencies would be expected to have the greatest effect on wells that produce water from just below the water table, because the more saline water would be entering at the water table and should reach these wells first and be least diluted by mixing. Inspection of the data reveals no relationship between increase in dissolved solids and depth of the bottom of the well below the water table. Indeed, well number D1-3-16aa, which has the largest increase (150 ppm) of any resampled well, is reported to be 244 feet deep with its water level at 190 feet. The depth to the bottom of the casing is not known. The next largest increase in dissolved solids in a resampled well was found in number D2-5-22ccd. The increase was 116 ppm, and the well produces from perforations between 90 and 165 feet while the water level is at about 13 feet (WSP 1482, p. 268). This lack of correspondence between rise in dissolved solids and depth below water table weakens the hypothesis that the probable increase in dissolved solids is related to the recent water shortage. However the hypothesis cannot be adequately evaluated without sampling the wells again next year.

With regard to hypothesis number 5, irrigation is known to cause increases in dissolved solids in groundwater when highly soluble minerals are present in the subsurface materials. Hem (1970, p. 166) summarizes the effect of irrigation as follows:

"When an area of low rainfall and accumulated solutes is reclaimed by irrigation, the increased water supply tends to leach away the solutes, and they appear in drainage water or return flow. The process is an acceleration of natural leaching and will increase the dissolved solids concentrations and loads in the residual water of the affected area for a considerable period."

Hem (p. 73) illustrates this effect with two irrigation wells in southwestern Arizona in which dissolved solids increased greatly over a period of 19 years. However, the increases in the Gallatin Valley are not limited to irrigated areas; and the greatest increase was at D1-3-16aa, which is in a dry land farming area. Consequently, it is doubtful if irrigation is the sole source of any general increase.

With regard to hypothesis number 6, the mechanism for increasing dissolved solids is similar to irrigation leaching, but the increased infiltration of water for leaching is a result of dry land agricultural practices rather than bringing in irrigation water. Such leaching is well known in areas elsewhere in Montana where saline seeps have developed. However, dissolved solids concentrations in the groundwater of saline seep areas are commonly much higher than they are in the Gallatin Valley. Evidence that this mechanism contributes to dissolved solids in the Gallatin Valley well be presented later in connection with the large increase found in well D1-3-16aa, but it cannot be the sole mechanism because dissolved solids also increased in wells located in irrigated areas as well.

With regard to hypothesis number 7, it is conceivable that changes in farming practices could cause small increases in dissolved solids. The solubility of carbonate minerals is greatly affected by the carbon dioxide concentration in the soil water. This concentration, in return, is affected by the amount of living organisms in the soil that produce carbon dioxide and the rate at which carbon dioxide can build up in the atmosphere of a particular soil. These things would seem to be related to the amount of nutrients in the soil, the air permeability of the soil, and probably numerous other factors as well. Such soil properties may change in accordance with changes in agricultural practices, and their effect on groundwater quality is not easily predicted. Although there have probably been many changes in farming practices since 1953, they are difficult to document. One change which has been documented is fertilizer use on wheat, which has been increasing in number of acres fertilized and average rate of application per acre in Gallatin County (Heid and Larson, 1974). Fertilizer, however, would not account for the increases in magnesium found in many of the wells. Consequently, it is probably not a major factor.

Hypothesis number 8 seems to be the best choice. It does seem likely that if the increase in dissolved solids is real, then in the dry land farming areas the increase is caused by increased leaching that results from more water moving to the water table than in pre-agricultural times. Likewise, in the irrigated areas, a strong possibility is that leaching by irrigation water is largely responsible for any general increase in dissolved solids. These two mechanisms differ primarily in the origin of the extra water for leaching and the details of the soil moisture regimen. Consequently, it might be said that a likely cause of the increase in average dissolved solids concentration is farming practices that increase the rate of leaching of subsurface minerals compared to pre-agricultural times. This is not to say, however, that septic tanks and recent changes in agricultural practices, especially fertilizer use, have no effect; but septic tanks do not seem to be important causes of the average increase in dissolved solids concentration found by the present data.

Individual Constituents

The increase in total dissolved solids is reflected in increases in several of the ion species. The average concentration of the various species are shown in Table 2. The mean concentrations labeled "all wells" are the means from all wells sampled in the 1951-53 and 1977 in their respective columns. These means exclude the wells that were not pumped before sampling in 1977. The wells excluded are A1-4-19cb and the resampled well at D2-5-15aa; an alternate well at D2-5-15aa was pumped and sampled, and it is included in the table. The mean concentrations labeled "resampled wells" are the means for samples taken from wells that were actually resampled in 1977, and all wells that had to be replaced by alternates in 1977 are not included. Resampled well number D1-5-22cd was also excluded because it has been deepened.

All of the constituents show increases in the resampled wells except sodium, chloride and fluoride. The increases in nitrate and sulfate are noteworthy because they have some bearing on human health. The nitrate increase was tested statistically, and the tests provided no support for the hypothesis that the difference in means represents an actual general rise in nitrate in the area. There is too much variance in the data and the sample is too small to allow that conclusion to be accepted with confidence. However, the data does suggest an increase that is in accordance with the increase in dissolved solids; see Figure A-13. The nitrate concentration in all samples analyzed in 1977 was well below the limit of 10 mg/l nitrate-nitrogen (approximately 44 ppm nitrate) that is applied to public water systems. However, the current average rate of increase (if any) is not known.

Mean concentrations Gallatin valley groundwater

Individual Wells

In addition to analyzing the valley-wide implications of the chemical analyses, the data was searched for indications of extraordinary changes in the vicinity of the individual wells that were sampled in 1951-53. This was done by looking for extreme differences between the 1951-53 values and 1977 values for the various constituents. First, differences in values obtained from resampled wells were analyzed. The procedure for each constituent was to compute the difference between the 1951-53 value and the 1977 value for each well. These differences were then ranked in order of difference with the greatest decrease in value at one end of the rank and the greatest increase at the other end. The greatest difference was tested to find whether it was significantly different from the other values. If it was significantly different, the next greatest difference was tested and so on. The test used is described by Dixon and Massey (1951, p. 275-276). When the test statistic was greater than the 90th percentile in Dixon's table of critical values, an effort was made to determine the cause of the difference between values for 1951-53 and 1977. This effort included a field inspection of the vicinity of the wells; and, in three cases, the owner of the well was contacted regarding possible causes of the difference. In all cases but one, the difference was consistent with causes that did not necessarily involve actual changes in concentration of constituents in the groundwater. These causes included (1) a well that had been deepened but was not reported as deepened during the original field investigation, and (2) a well that was classified as a resampled well but was really an alternate. Differences in potassium were not investigated because (1) the largest difference was only 4.15 ppm except for one sample that contained water from a well that could not be pumped and (2) potassium differences did not appear to be normally distributed when plotted on normal probability paper, and the statistical method requires the data to come from a normally distributed population. The one case where differences in constituents seem to represent an extraordinary change in the chemical composition of the water is the change in composition of the water from well number D1-3-16aa.

In this well the statistical tests indicated extraordinary rises in total dissolved solids, sulfate, nitrate, and boron. Dissolved solids increased from 290 ppm to 440 ppm, sulfate increased from 65 ppm to 165 ppm, nitrate increased from 4.2 ppm to 14.0 ppm, and boron increased from 0.12 ppm to 0.22 ppm. The well is located in a farming area about 1 1/2 miles west of Amsterdam. It was reported as not having been deepened, a septic system has not been installed near the well since 1951, and the land near the well is not irrigated. The amount of fertilizer applied to land in the vicinity of the well has increased, but since magnesium increased from 8.4 ppm to 17.64 ppm it is doubtful if fertilizer is primarily responsible for the increases. A possible mechanism is that farming practices have increased the amount of water percolating to the water table in the area, and the associated leaching of minerals is causing a gradual increase in dissolved solids. It is beyond the scope of the present investigation to attempt to predict the future quality of water at this location. However, if the increase in nitrate is real, and if it continues long enough, 44 ppm will eventually be reached, and there will be some cause for concern about the health effects of the water . Also, possible future increase in sulfate could eventually produce water that would have effect on some people not accustomed to drinking it. The reader is referred to Comely (1945), Keller and Smith (1967), Walton (1951), and Young et al (1976) for information on nitrate hazard.

SUMMARY AND CONCLUSIONS

The information in U. S. Geological Survey Water Supply Paper 1482 on groundwater levels and groundwater chemistry in the Gallatin Valley was checked and updated by remeasuring water levels in wells and by resampling wells for chemical analysis. In addition, alternate wells were measured in many cases where the original well was no longer available for water level measurements, and alternate wells were sampled in most of the cases where water samples could not be obtained from the original well with a reasonable amount of effort.

No evidence for long-term change in water levels was found, and the water level map and depth to water map in WSP 1482 are not significantly changed by the 1977 data. With regard to temporary change, water levels in the valley were generally lower in 1977 than in 1951-53, and they were considerably lower in the vicinity of Belgrade. These lower water levels are attributed to an abnormally small amount of irrigation water available in 1977 and less spring infiltration of water in the dry land farming areas.

The new data on water chemistry indicates that the average of total dissolved solids concentrations in wells of the Gallatin Valley area has probably increased. This increase, if it is real, is attributed primarily to increased leaching of minerals in the subsurface caused by increased percolation of water through the soil to the water table that began with the change from uncultivated land to irrigation and dry land farming that has been in progress for over a century. However, other factors might also contribute to an increase.

At present the general quality of the groundwater is good, and the probable average rate of increase in dissolved solids is so small that it may be expected to remain good in the near future. The prognosis for the distant future in uncertain.

With respect to local water quality problems near the sampled wells, only one well showed an extraordinary increase in any constituents. This well is in a farming area near Amsterdam, and it is probably just a more severe case of increase in dissolved solids related to leaching of minerals in the subsurface. Nitrate in this well rose from 4.2 ppm to 14 ppm and the prognosis is uncertain.

REFERENCES

Comely, H. H. (1945): Cyanosis in infants caused by nitrate in well water; Jour. American Medical Association: vol. 129, p. 112.

Davis, S. N. and R. J. M. DeWiest (1966): Hydrogeology; John Wiley and Sons.

Dixon, W. J. and J. J. Massey, Jr. (1957): Introduction to Statistical Analysis, Second Edition: McGraw-Hill.

Environmental Protection Agency (1976): Manual of Methods for Chemical Analysis of Water and Wastes; Report No. EPA-625-16-74003A, Environmental Monitoring and Support Laboratory, Environmental Research Center, Cincinnati, Ohio, 45268.

Hackett, O. M., F. N. Visher, R. O. McMurtrey and W. L. Steinhilber (1960): Geology and ground water resources of the Gallatin Valley, Gallatin County, Montana; U. S. Geological Survey Water Supply Paper 1482.

Heid, W. G. and D. K. Larson (1974): Fertilizer use in Montana; Montana Agricultural Experiment Station Bulletin 628 (Revised).

Hem, J. D. (1970): Study and interpretation of the chemical characteristics of natural water, second edition; U. S. Geological Survey Water Supply Paper 1473.

Keller, W. D. and G. E. Smith (1967): Ground-water contamination by dissolved nitrate; Geological Society of America Special Paper 90, P. 47-59.

Montana State Engineer's Office (1953): Water Resources Survey, Gallatin County, Montana; Available from Montana Department of Natural Resources and Conservation, Water Rights Bureau.

Walton, G. (1951): Survey of literature relating infant methaemoglobinemia due to nitrate contaminated water; American Jour. Public Health, vol. 41, p. 986.

Walton, W. C. (1970): Groundwater Resource Evaluation; McGraw-Hill.

Young, C. P., D. B. Oakes and W. B Wilkinson (1976): Prediction of future nitrate concentrations in ground water ; Ground Water, vol. 14, p. 426-438.

ACKNOWLEDGMENTS

Many residents of the Gallatin Valley contributed to this investigation. They granted permission to measure water levels in their wells and allowed samples of well water to be collected for chemical analysis. The U. S. Geological Survey, Water Resources Division, Helena, Montana supplied considerable data and information including detailed information on the water wells included in Water Supply Paper 1482.

APPENDIX

Gallatin Valley Montana groundwater level, 1977

Gallatin Valley Montana precipitation graphs

Gallatin Valley Montana surface water inflow and outflow

ESTIMATED DRAWDOWN BY BELGRADE MUNICIPAL WELLS

Belgrade obtains its water from two municipal wells. If the population served is estimated at 1800 people; and if the average per capita use is 180 gallons per day (gpd); then the average use would be 324,000 gpd, which will be conservatively rounded to an estimated use of 325,000 gpd. This use would require an average pumping rate of about 226 gallons per minute (gpm).

WSP 1482 gives a transmissivity (T) of 240,000 gpd/ft for a Belgrade municipal well (D1-4-1dc), and it also gives 0.15 for an average specific yield for valley-fill material in the Gallatin Valley. The delay index for gravel is probably about 1 (Walton, 1970, p. 245). These values will be used to estimate the drawdown near a Belgrade water well if it were pumped at 226 gpm for 25 years. The actual drawdown at either of the pumping wells would be less than the amount thus calculated if the assumptions of the Boulton equation (Walton, 1970, p. 222) were met and the proper values for T and Sy have been selected.

Delayed yield from storage should not be significant after 25 years. For example, at a radius of 2 feet and time twt from start of pumping at which the effects of delayed yield are negligible is less than 3 minutes. This value of twt was obtained from Walton (1970, Figure 4.31). At larger radii, twt is still less than 25 years, and the Theis equation may be used with Sy in the position of the storage coefficient.

The Jacob approximation of the Theis equation may be used for distances (r) of less than 8836 feet since uy = (2693*r2*Sy)/(T*t) = 0.01 at this distance (t is time since pumping started in minutes and Sy is specific yield). The Jacob equation is

s = ((264*Q)/T)*log10((0.3*T*t))/(r2*Sy)),

where s is drawdown, Q is pumping rate in gpm, r is radial distance from the center of the pumping well (ft), and t is in days. When r is 2 feet, this equation yields 2.2 feet as the drawdown after 25 years of pumping. At greater distances from the well the calculated drawdown, of course, would be less. Even if the estimates of T and Sy were considerably in error, the municipal wells would still be unlikely to produce an extensive cone of depression. An analysis of the small amount of industrial pumping near Belgrade would yield similar results.