Abstract

During 1975-85 major advances in understanding the effects of agricultural land-use on ground-water quality have been achieved. These advances are briefly reviewed and illustrated mainly by results of British research. Marked improvements in sampling methods, especially deeper in the vadose zone, and in analytical techniques, notably for synthetic organic compounds and environmental isotopes, have revealed the profound, but delayed, impact produced by changes in agronomic practices. Elevated rates of nitrate leaching from well-drained soils have been demonstrated for a wide range of climatic types and agricultural systems. Moreover, increasing irrigation efficiency on such soils is causing some serious ground-water salinization problems. Certain of the more mobile pesticides have proved far more persistent in the subsurface than might have been believed possible from their observed degradation in fertile soils. This causes serious preoccupation about future prospects for ground-water quality. Environmental administrators have been slower to confront diffuse-source pollution from agricultural land-use than industrial point-source pollution problems. Control measures to protect potable ground-water sources are thus discussed.

General Background

Trends in Agricultural Land Use

In recent decades there has been a radical evolution in agronomic practice in many regions of the world associated with (largely successful) attempts to increase agricultural productivity. The intensification of production from agricultural land is being sustained by the application of ever-increasing quantities of inorganic fertilizers and a wide spectrum of synthetic pesticides.

In the more arid regions, the frequency of cropping has beep increased and additional land has been brought into production through new irrigation schemes and increasing irrigation efficiency. Some of these schemes are utilizing water of inferior quality (including wastewater, although this aspect is not con sidered further here). Areas of sandy, well-drained, soil have been purposefully selected in such instances because they are less prone to waterlogging and salinization.

All these changes were initiated in the industrialized nations during the 1950's and have spread, or are spreading, to many developing nations. Common trends include the replacement of traditional crop rotations by intensive and continuous cultivation of high-value crops selected according to prevailing market, soil and climatic conditions. In many instances, near monocultures across extensive tracts of agricultural land have resulted. Examples include cereal grain belts, sugar, coffee and other plantations, market gardening centers and citrus orchards.

Nature of Diffuse Ground-Water Pollution Risk

In many nations the principal recharge areas of numerous aquifers form valuable tracts of farming land and are now almost completely used for intensive crop cultivation. In such cases the bulk of the replenishable ground-water resources originate as excess rainfall and excess irrigation infiltrating this land. In consequence, these resources are highly susceptible to contamination by agronomic practices. The large areal extent of agricultural activities makes the potential for diffuse pollution all the more serious.

There is significant risk of elevated rates of nutrient, salt and pesticide leaching to ground water from cultivated soils (Vrba & Romijn, 1986) with the corresponding potable water-quality guidelines (Table 1) being exceeded. This is especially the case in the areas of well-drained (thin and/or sandy) soils widely found in aquifer recharge areas to which this paper principally refers.

*range for individual listed insecticides or herbicides, but many remain to be evaluated because of lack of medical evidence; to date the insecticides - aldicarb, DBCP, EDB - and the herbicides - atrazine, 2.4D, MCPP- have been found most widely in ground water
**does not consider Se, B, As, F, etc. which may also be leached from some irrigated soils and from the vadose zone
** for most sensitive crops, many others can tolerate 5-10 times higher concentrations than those indicated

Table 1
Summary of Water-Quality Guidelines related to the Contamination
of Ground Water by Agronomic Practices

Past Complacency about Ground-Water Quality Deterioration

Prior to the mid-1970's there was widespread complacency about such risks. Environmental and agricultural sector administrators have been slow to realize, or to accept, the scale of potential problems. This complacency can be attributed to a number of contributory causes:

(a) slow average rates of vadose zone transport of contaminants leached from cultivated soils (Foster, 1976;
      Young et al, 1976; Bouwe~, 1987), at maximum normally not exceeding 3 m/a and 8 m/a beneath
      non-irrigated and irrigated land respectively,
(b) the resultant delayed impact (more of a legacy) for ground-water quality, which is most pronounced in deep
      water-table and/or overpumped aquifers (Bouwer, 1987),
(c) lack of conciousness among ground-water specialists about the scale and relevance of changes in
     agriculture land-use practices,
(d) past limitations on the sampling of water from deep in the vadose zone,
(e) preoccupation in the agricultural sector with problems seriously reducing agricultural productivity, such as
      waterlogging and soil salinity due to rising ground-water tables, but not with ground-water quality
      deterioration,
(f) most agricultural research stations are sited on soils requiring artificial drainage (because of the interest in
      monitoring drainage waters), and these are characterized by higher rates of denitrification and stronger
      sorption of pesticides than are well-drained soils.

Impact of Non-Irrigated Arable Cultivation

The greatly increased rate of nitrate leaching resulting from the conversion of pasture land to continuous cereal cultivation is dramatically illustrated in Figure 1. These vadose zone profiles were obtained by microanalysis of porewater centrifuged from core-drilled samples. Although interpretation of such data is complicated by dispersion during vertical unsaturated flow, the depth of penetration of ground-water contaminants is consistent with the date of land-use change and the environmental tritium distribution for same site.

Figure 1
Vadose zone profiles from the chalk of east Yorkshire-Britain
showing the effect of changing Non-Irrigated Agricultural Land Use
(modified after Foster et al, 1982)

It was concluded (Foster et al, 1986) that oxidation of organic nitrogen is promoted and denitrification suppressed in light soils well aerated by cultivation practices. Soil nitrate is then readily leached by excess rainfall at times when present in excess of plant requirements. In contrast, the soil compaction and root system under pasture land lead to greater nitrate uptake and to significant soil denitrification losses, even from well-drained soils. However, organic nitrogen can accumulate, be oxidized and then leached in large quantities following ploughing and reseeding (Young et al, 1976).

The former conclusion was corroborated by numerous vadose zone profiles beneath long-standing arable land with similar soil in a somewhat drier area of Britain (Figure 2). The figure also shows that intensification of arable production has resulted in increasing nitrate leaching losses since the mid- 1960's, with deep percolation containing nitrate much in excess of 50 mg N03/1. This probably results primarily from augmentation of the soil organic nitrogen pool, consequent upon the larger weight of crop residues associated with increased production, and secondarily from direct leaching of fertilizer nitrate. Elevated sulphate concentrations in the same profiles, with peaks at depths of 5-15 m bgl (at the date of sampling) appear to reflect the large quantities of sulphate in N-P-K fertilizers, especially in the years before ammonium nitrate replaced ammonium sulphate as the main source of fertilizer nitrogen. Nitrate leaching from arable land has been shown to very considerably iwith certain facets of arable management practice, including the rate and timing of nitrogen fertilizer applications (Foster et al, 1986).

Figure 2
Vadose zone profiles from beneath Non-Irrigated Long-Standing Arable
Land on the chalk of west norfolk-britain (modified after Foster & Young, 1980)

Sensitivity of Nitrate Leaching to Pasture Land Management

Since pasture land is less prone than cultivated land to nitrate leaching, it would appear to offer a useful option for controlling aquifer nitrate pollution. However, more detailed investigation (Figure 3) has demonstrated that leaching rates from pasture on well-drained soils increase abruptly and unpredictably to elevated levels when grassland productivity is intensified by heavy applications of nitrogen fertilizer and high density grazing (Foster et al, 1986). This phenomena appears to require more detailed study in a variety of climatic regimes and soil types before a general conclusion can be reached.

Irrigated Cultivation: Theoretical Possibilities and Practical Realities

Where irrigation is practised, there exists the possibility of controlling soil moisture so as to maximize nutrient uptake and to restrict deep percolation, thereby controlling the leaching of nitrates and pesticides. This is most practicable where virtually all plant moisture requirements are provided by irrigation and where the wet season is confined essentially within 2-3 months of each year. It is less feasible where only supplementary irrigation is required to secure a primarily rainfed crop, but even then the maximization of nitrate uptake can be assured by providing optimum moisture levels at times of rapid plant growth, and thereby reducing soil nutrient residues. Moreover, denitrification losses rather than nitrate leaching may become more significant with irrigation or wetter climates, especially for finer-grained soils.

With some notable exceptions, however, irrigation practices remain relatively inefficient and considerable excess moisture is applied with each irrigation lamina. Regular deep percolation and leaching of nitrates and pesticides results, especially from well-drained soils. This has been judged the cause of steadily-increasing concentrations of ground-water nitrate ín various regions of the world (Saffigna & Keeney, 1977; Gormley & Spalding, 1979; Panda, 1979; Sabol et al, 1987; Schmidt & Sherman, 1987; Beek et al, 1988).

Where highly-efficient irrigation techniques have been introduced, greater control over leaching is possible, but ground-water recharge will become progressively more saline (Bouwer, 1987). Nutrient leaching 'will be much reduced, but nitrate concentrations may remain high because of the much smaller volume of deep percolation. Controlled precipitation of CaC03 and CaS04 in the vadose zone is theoretically possible.

Field evidence, however, suggests that with increasing irrigation efficiency salinization of ground-water recharge will cause severe problems in numerous arid regions, especially those where ground water itself is the only source of irrigation (Figure 4). Problems should be less pronounced where low-salinity surface water constitutes the bulk of irrigation supply, where seepage from unlined irrigation canals is an important source of ground-water recharge and where regional ground-water flow from uncultivated areas occurs. However, the deep percolation generated by the initial irrigation lamina used to habilitate saline soils for cultivation is also believed to have contributed to the increase of ground-water salinity in arid regions, such as parts of Sonora, Mexico.

Figure 4
Salt and Nutrient leaching to ground water beneath high-efficiency
Irrigated cultivation in the harquahala Valley-Arizona
(data from Sabol et al, 1987)

Surface Persistence of Pesticides

Prior to 1975, there was not much concern about the possibility of ground-water pollution by pesticides, despite the fact all such compounds ark, to greater or lesser degree, chemically tailored to be both toxic and persistent. Agricultural scientists argued that soil sorption of the higher molecular weight compounds (such as the chlorinated hydrocarbon insecticides) and ~'olatilization of lower molecular weight compounds (like most herbicides) would predominate.

Since in fertile soil most pesticide compounds have half-lives of less than 1 year and many of less than 2 months, it was assumed that soil residues would be eliminated by anaerobic biodegradation or chemical hydrolysis, thus preventing ground-water contamination, except possibly where the water-table was shallow, soils coarse grained, and pesticide applications excessive.

There is considerable uncertainty about permissible concentrations of pesticides in potable water supplies. The EEC drinking water directive sets admissible concentration for any individual pesticide compound of whilst the WHO and US-EPA make specific recommendations for some compounds allowing higher concentrations in most cases (Table 1).

Since the late 1970's there have been reports of an increasing pesticides penetrating into various aquifers used for potable water some cases at concentrations giving serious concern for human health al, 1984; Lawrence & Foster, 1987). Concentrations in excess of 10 tug/1 of aldicarb, a soil insecticide used mainly in potato production, resulted in closure of water-supply boreholes and in restriction on future applications in Long Island, New York (Zaki et al, 1982) and on the Great Sand Plain in Wisconsin (Rothschild et al, 1982). Two other soil insecticides, DBCP and EDB, formerly used widely in irrigated fruit-growing areas, have been detected in troublesome concentrations (above 1 u g/1) in Arizona (Sabol et al, 1987), California (Schmidt & Sherman, 1987) and Hawaii (Oki & Giambelluca, 1987).

Atrazine, used as a herbicide for irrigated maize production, has been widely detected in ground waters of Nebraska (Wehtje et al, 1981) and Iowa;(Hallberg et al, 1983) at concentrations in excess of 1 ug/1, although this level is not believed to represent a risk to potable water supply or a problem for irrigation use. In Britain, the common herbicides MCPP, 2,4-D and atrazine have been reported individually at concentrations up to 2 ug/1 in some public water-supply boreholes in arable farming areas on the East Anglian Chalk (Croll, 1985).

All of the above pesticides are among the more mobile, and thus most readily leached from cultivated soils, although losses of no more than 5%, and generally less than 1%, of the applied dose are normally involved. While their quoted half-lives are relatively short, these apply only to fertile soils and should not be taken as representative of thin and/or sandy soils widely present on aquifer outcrops, and even less of the vadose zone where microbial activity and organic carbon content are very much lower (Bouwer, 1987; Lawrence & Foster, 1987). Subsurface transport of most pesticides will be retarded due to partition between a mobile aqueous phase and an immobile sorbed phase. There is, however, also the possibility in coarse-grained or fissured aquifers of transport of pesticides sorbed onto colloidal particles (McDowell-Boyer et al, 1986) and the extent of this phenomenon requires further research.

To date, problems of obtaining unbiased samples of adequate volume have precluded the determination of depth profiles for pesticide compounds in vadose zone pore water to determine the scale of any attenuation. This is a key area for future research. For the present, it would be prudent to assume that no degradation occurs once a pesticide has been leached below the base of the root zone. Evidence from a saturated zone profile in an aquifer of the San Joaquin Valley, California, in the irrigated, sandy, citrus-growing area around Fresno (Schmidt & Sherman, 1987), shows a close correlation between the penetration of nitrate (in the range 5-25 mg N03-N/1) and DBCP (in the range 5-20 jig/1). It appears that there is little attenuation of either.

Concluding discussion

Future Outlook for Ground-Water Quality

The medium-term prospect for ground-water quality in respect of diffuse groundwater pollution is not promising, especially in areas of well-drained soils. This is mainly because of the legacy of high leaching rates of nitrates and some pesticides in past agrononomic practices. This legacy takes the form of elevated concentrations of pollutants already in slow percolation through the vadose zone or already contaminating the upper parts of aquifers but still not affecting water-supply boreholes.

In individual cases, the scale of future problem will depend on what percentage of the total aquifer recharge area has been under cultivation súbject to high leaching losses, on the thickness of the vadose zone, on the extent of any in-situ contaminant degradation/elimination within the ground-water system and on which of the previously-applied pesticides were sufficiently mobile to be leached.

In numerous aquifers it will be impossible to avoid the abandonment of some potable water-supply boreholes, or the development of new sources for blending or the installation of costly and problematic treatment plants.

Options for Ground-Water Pollution Control

In the longer run, measures to control nitrate and pesticide leaching are required to protect potable ground-water supplies or at the very least to bring about more stability in ground-water quality to allow effective planning in the water-supply industry. It would be highly desirable for the entire recharge areas of important aquifers to be regarded as a special case when defining best agronomic management practice, taking into account possible ground-water quality deterioration. In particular, it is essential to discourage and to penalize any practices that run high risk of adverse impact on ground-water quality, while offering only very marginal economic return to the agriculturalist.

The most positive policy is to take even closer control over land use in so-called wellhead protection zones around major potable water-supply sources. These two options should not be regarded as mutually exclusive since the more successful is the control over the entire recharge area the smaller can be the wellhead protection area.

Hydrogeological research during the period 1975-85 has identified most of the factors influencing the scale of impact of agronomic practices on ground-water quality (Table 2) and, therefore, those measures which will help control the problem.

* effect on soil leachate concentration not on soil leaching load, since for latter recharge volume must also be
   taken into consideration
**where applicable
(+)+ tends to increase concentration
0      minimal effect
(-)-   tends to decrease concentration

Table 2
Simplified summary of the relative impact of key agronomic
factors upon ground-water quality

In wellhead protection zones, the following measures would appear generally to be the most appropriate:

(a) selection of crop types to ensure maximum continuity of plant cover,
(b) restriction on the amount of ploughing undertaken in any single year,
(c) limiting grazing densities,
(d) controlling the timing and amount of nitrogen fertilizer applications,
(e) controlling pesticide applications, including the banning of any exceptionally hazardous compounds of
     significant mobility in sandy soils.

In certain cases it may be decided desirable not to permit any agricultural activity in the wellhead protection zone. In such instances, low productivity grassland and woodland would generally be the most acceptable, including use for public recreation.

In areas where agricultural cultivation is dependent upon irrigation, the most attractive option to control nutrient and pesticide leaching to ground water unquestionably will be maximizing irrigation efficiency and restricting periods of soil leaching to those with low nutrient and pesticide residues in the soil. However, the inevitable effect of this policy will be to reduce the volume of ground-water recharge and to greatly increase its salinity. The associated salinization of ground water can be delayed, and perhaps somewhat reduced, but ultimately removal of saline recharge may be necessary if aquifers are not to suffer irreversible deterioration.

Acknowledgements

The author is grateful to the Director of the Pan American Health Organization for permission and for support to present this paper. The views expressed are those of the author and not necessarily of the Pan American Health Organization.

The author's interest in this subject was developed, and much field data generated, through a research program of the British Geological Survey, which was funded by the (British) Department of the Environment, the European Economic Community and the (British) Natural Environment Research Council. He wishes to acknowledge the contribution of many colleagues to this program, especially Adrian Lawrence and Judy Parker.

References

Beek C G E M van, Boukes H, Rysbergen D van, Straatman R. 1988. The threat to Netherlands waterworks of nitrate in abstracted groundwater as demonstrated by the Vierlingsbeek wellfield. Water Supply 6(3):313-318.

Bouwer H. 1987. Effect of irrigated agriculture on groundwater. J Irrig Drain Eng 113:4-15.

Cohen S Z, Creeper S M, Carsel R F & Enfield C G. 1984. Potential for pesticide contamination of ground water resulting from agricultural uses. Amer Chem Soc Sym Series 259:297-325.

Croll B T. 1985. The effects of the agricultural use of herbicides on fresh water. Proc WRC Conf "Effects of Land Use on Fresh Water" (Stirling, June 1985):13.

Foster S S D. 1976. The vulnerability of British groundwater resources to pollution by agricultural leachates. MAFF Tech Bull 32:68-91.

Foster S S D Bridge L R Geake A K, Lawrence A R & Parker J M. 1986.1 The groundwater nitrate problem. BGS Hydrogeology Report 86-2.

Foster S S D, Cripps A C & Smith-Carington A K. 1982. Nitrate leaching to groundwater. Phil Trans Royal Soc London B: 296:477-489.

Foster S S D & Young C P. 1980. Groundwater contamination due to agricultural land-use practices in the United Kingdom. Hydrol Series 30:268-282.

Gormly J R & Spalding R F. 1979. Sources and concentrations of nitrate-nitrogen in groundwater of the Central Platte Region, Nebraska. Ground Water 17:291-301.

Hallberg G R, Hoyer B E, Bettis E A & Libra R D. 1983. Hydrogeology, water quality and land management in the Bay Spring Basin, Clayton County, Iowa. Iowa GS Open-File Report 83-3.

Handa B K. 1979. Effect of return~flows from irrigated lands on the chemical composition of ground water from shallow unconfined aquifers. Tech 11:337-349.

Lawrence A R & Foster S S D. 1987. The pollution threat from agricultural pesticides and industrial solvents. BGS Hydrogeology Report 87-2.

McDowell-Boyer L M, Hunt J R & Sitar N. 1986. Particle transport through porous media. Water Resources Research 22:1901-1921.

Oki D S & Giambelluca T W. 1987. DBCP, EDB and TCP contamination of ground water in Hawaii. Ground Water 25:693-702.

Rothschild E R, Manser R T & Anderson M P. 1982. Investigations of aldicarb in ground water in selected areas of the Central Sand Plain in Wisconsin. Ground Water 20:437-445.

Sabol G V, Bouwer H & Wierenga P J. 1987. Irrigation effects in Arizona and New Mexico. J Irrig Drain Eng 113:30-48.

Saffigna P G & Keeney D R. 1977. Nitrate and chloride in ground water under irrigated agriculture in Central Wisconsin. Ground Water 15:70-177.

Schmidt K D & Sherman I. 1987. Effect of irrigation on groundwater quality in California. J Irrig Drain Eng 113:16-29.

Vrba J & Romijn E. 1986. Impact of agricultural activities on ground water. IAH Int Contrib Hydrogeology 5.

Wehtje G, Leavitt J R C, Spalding R F, Mielke L R & Schepers J S. 1981. Atrazine contamination of groundwater in the Platte Valley of Nebraska from non-point sources. Studies Environ Sci 17:141-145.

Young C P, Oakes D B & Wilkinson W B. 1976. Prediction of future nitrate concentrations in ground water. Ground Water 14:426-438.

Zaki M H, Moran D & Harris D. 1982. Pesticides in groundwater: the aldicarb story in Suffolk County, New York. Amer J Public Health 72:1391-1395.

Comentarios al Webmaster

[ Homepage CEPIS   ]