Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

8. Swamp transformation in the Sanjiang plain by the well-drainage and well-irrigation

Zeng Jianpiny, Wang Chunhe. and Sun Guangyou
Institute of Geography, Chinese Academy of Sciences, Changchun, Jilin Province


The Sanjiang Plain is the largest swamp area in China. Owing to its low and level relief. clay soils and poor drainage, about 2 million ha of the swampland have not yet been put to full economic use. The area is threatened by frequent waterlogging and drought.A wellpumping system that combines drainage with irrigation, and which can be effectively used to develop the swamp land for agriculture, has been successfully designed.


The Sanjiang Plain contains the largest area of wetland in China, but owing to the low-lying, flat relief, heavy clay soils and poor drainage, about 1.98 million ha of these wetlands have not been put to full use, and the existing cultivated land is frequently threatened by waterlogging. Although open-ditch drainage is effective where there are good drainage outlets, it loses much of its effectiveness in the vast interfluves of this region, where drainage is poor or where nearby discharge areas are lacking. Further, the open-ditch method is useful only for drainage and cannot solve the water problem in the Sanjiang Plain, where spring and autumn rainfalls differ greatly, and seasons of drought and waterlogging alternate. It is imperative, therefore, to devise a way to drain excess swamp-water which is not linked with the groundwater, and to reserve it for irrigation during the dry season. This would ensure higher and more stable crop yields regardless of drought and waterlogging, and thus would contribute to the economic development of the Sanjiang Plain.

In 1974, we began a research project on the transformation of swampland by well drainage and well irrigation. Since then, three experimental stations have been established. The first, located near the headquarters of the Qianjin State Farm, experimented with single-well drainage and showed the feasibility of using wells to drain areas with micro-pressure water in the central part of the Sanjiang Plain. The second station, on the Sixth Branch Farm of the Qixiang State Farm, experimented with the potential benefits of well drainage and well irrigation. There the control of waterlogging by well drainage and irrigation by spraying pumped water produced bumper crops of wheat and soybeans in the same year, thereby proving that the method can be used both to drain swamps and irrigate the reclaimed land to produce higher crop yields. In order to establish the feasibility of using a group of wells for drainage and irrigation over a large area, regional hydrogeological investigations were conducted by the third experimental station, in Team Four of the Qianjin State Farm, during August 1976.

The results of the experiments indicate that the well drainage and well-irrigation method is not hampered by the discharge conditions on the surface and that the same well can be used for drainage or irrigation as circumstances require. Other advantages of the method include a small space requirement, modest investment and quick results. It can also promote the interchange of groundwater with surface water. Thus the method could be essential to the successful agricultural modernization of the Sanjiang Plain.

Regional Hydrogeology

The Sanjiang Plain is low-lying and swampy with elevations of 50-70 m above sea-level. Relief dips gently from south-west to north-east, with a gradient of about 1 :10,000. Geomorphologically the plain is characterized by first-grade terraces and alluvial flats, and the Quaternary system consists mainly of alluvial deposits with an average depth of 120-200 m and a maximum depth of 240-280 m. The surface soils consist of sandy loam, loam, and clay soil, with a thickness of between several to more than 10 m. Below is an interbedding of gravelly sand and sandy gravels intercalated with a small amounts of clay soils, and together they form a thick, complex and lenticular Quaternary aquifer. The base consists of Neogenic mudstone and pelitic siltstone. The aquifer is rich in water and quite permeable, with a coefficient of permeability of 10-120 m/day. The specific yield per well is 7.2100.8 m/hr-m.

Fig. 8.1. Hydrogeological Profile from the River Jiansan to the River Tong

There are obvious differences in the surface materials of the western and eastern parts of the plain. In the eastern half, except for the three large alluvial flat areas, the surface is covered by brownish-yellow and brown clay or loam with a thickness of 3-17 m. The low permeability (the coefficient is 0.0013-0.635 cm/day) effectively prevents the surface water from contacting the groundwater. Therefore, the groundwater here has micropressure, and a steady watertable of 2-10 m. In the lower reaches of the rivers, where the topography is slightly more elevated, the groundwater is contained in inter-layer pores (fig. 8.1).

Thus the swamps in the eastern part of the plain are fed mainly by rainfall and surface runoff. In the western part and in the three major alluvial flat areas, where the surface materials are coarse, the clay soils are thin and they appear only locally. Sandy loam, sand, and even gravel strata often appear on the surface. The groundwater is contained in pore spaces at a depth of 1-4 m. In some low valley flats and in areas with ancient river courses, the water is less than 1 m below the surface or even on the surface, which then becomes a source of recharge water in this swamp area. The yearly magnitude of change in groundwater level is 1-1.5 m, except along the rivers, where it is 3-4 m.

The sources of groundwater recharge in the plain are precipitation, mountain streams, infiltrated floodwater, and floodwater returning from the River Songhua. The direction of flow of the groundwater for the whole region follows the trend of the land surface towards the Heilong, Wusuli and Songhua rivers.

The groundwater is mainly fresh with HCO3-Ca-Na and a low level of mineralization. In some places the water contains HCOCI-Ca and Mg. The degree of mineralization is 0.2-0.5 g/l, but in some localities it is 0.5-1 g/l. The pH value is 6-7, i.e. slightly acidic. The Ka values of coefficient of irrigation are all higher than 18. The iron anion content is high, commonly 1-5 mg/l. The water temperature is fairly low, between 4 and 8C.

Thus, the basic characteristics of the region's hydrogeology are that the aquifers are thick, at a shallow depth, very permeable, and rich in water, and that the water quality is good and the static water level ideal. The conditions are favourable for the popularization and utilization of the well-drainage and wellirrigation method.

The hydrogeologic conditions of the third test site are representative of those of the region as a whole (fig. 8.2). There are three visible layers in the Quaternary deposits, the total thickness of which is 199 m. The upper layer contains yellowish-brown clay and loam with a thickness of 6-10 m. The middle layer is the bluish-gray and fine silty sand layer, with a thickness of 5-10 m. The lower layer has grayish-white, rudaceous, and medium coarse sand and gravels. The groundwater has micro-pressure and its steady watertable is about 5 m below the surface. The flow direction is N 65E and the hydraulic gradient is 1 :7,000.

Fig. 8.2. Cross-section of the Hydrogeology at the Third Experimental Station

Fig. 8.3. Sketch Map of the Third Experimental Station, Qianjin State Farm, Heilongliang Province

Fig. 8.4. Q-T and S-T Curves for Three Pumping Drawdowns, Main Well No. 3, Third Experimental Station

Analysis of Test Results

The test sites total about 660 ha. Seventeen wells were set up for both drainage and irrigation. The well depth is about 42 m, and the diameter of most is 400 mm. All have permeable walls. Well no. 3 is the " Main Test Well, " around which are six observation wells (fig. 8.3). The components of the well system are simple, consisting of a power generator, pumps, valves and tubes.

Multiple-well Pumping Test

Based on the calculated data for fixed discharge with unsteady flow plus steady flow, we carried out three positive pumpings to examine drawdowns at Well No. 3. The results are given in table 8.1. Figure 8.4 shows that the unsteady flow lasts for a short time only before the stable point appears. The turning point of water-level recovery on the hydrograph is very clear, showing that the water returns quickly to a level close to the original one.

By the calculation of the first order difference and error estimate, we determined the Q = f (s) curve (fig. 8.5) was a parabola. The curve is away from the Saxis. The empirical parameters of the curvilinear equation may be calculated by the following analytical method :

After substituting it into the empirical parabolic equation, the yield from a single well when the drawdown level reaches 2 m was calculated as follows:

To calculate the coefficient of permeability (K) of steady flow, we applied the formula, c?? 0, c + 1 <0.5 m, which had been used for a partially artesian well with a very thick aquifer. This calculation is composed of formulae of two vertical observation wells:

The average coefficient of permeability of three drawdowns equals:
K= 65.04 m/day

Fig. 8.5. Q-S Curve of Pumping Test, Main Well No. 3

TABLE 8. 1 Results of Pumping Test, Well No.3 Third Experimental Station

Yield Specific yield Coefficient of permeability
m/h 33/day m/h/m Average m/day Average
S1 = 0.653 71.00 1,704.00 108.73   71.84  
S2 = 1.049 104.17 2,500.08 99.30 100.83 62.54 65.04
S3 = 1.447 136.69 3,280.56 94.46   60.74  

TABLE 8.2. Hydrogeological Parameters Determined during Pumping of Unsteady Flow

  Well 1 Well 5
Coefficient of water transmissibility ( T) (m/day) 12,514.84 12,632.77
Coefficient of pressure transmissibility (a) (m/day) 1.63 x 107 2.1 x 107
Elastic release water coefficient 7.7 x 10-4 6.02 x 10-4
Coefficient of permeability (K) (m/day) 65.70 64.22
Duration (minute) 25 25
Level of recovery    
water calculated (m) 0.046 0.044
water observed (m) 0.050 0.050
Difference (m) -0 004 -0.006

TABLE 8.3. Test Results of Multiple Well Drainage

of pipe
Raised water-level(m) Discharge Average
Average Maximum m/h m/day
G5 180 37.0 22 2.50 0.500 0.660 39.51 948.24 79.02
G4 300 39.0 22 2.50 0.417 0.445 41.40 993.60 99.28
  2.50 0.457 0.500 43.00 1,032.00 94.09
3 400 40.25 24 3.33 0.630 0.690 95.04 2,280.96 150.86
  8.33 1.748 2.375 144.02 3,456.48 82.39

TABLE 8.4. Hydrogeological Parameters during Pumping and Injection of Water

Form Well No. Water
(m /day)
Pumping G1 12,514.84 1.63 x 107 7.7 x 10-4 65.70 65.04a
Injecting G1 12,370.93 1.46 x 107 8.47 x 10-4 64.90 63.01

a Based on calculated data of wells G4 and G5

According to the pumping data of steady flow, when Q = 136.69 m/h, the value of the radius of influence, R, is:

and R= 295.4 m.

We adopted the grating method of the standard curve of the partially penetrated artesian well to calculate hydrogeologic parameters during the pumping of unsteady flow:

The calculation result is listed in table 8.2. We checked the level of water recovery by using the following formula :

The result shows that the calculated parameters correspond basically with the actual situation of the aquifer during the time of pumping.

According to the designed drawdown, the water yield of the unsteady flow for a single well is:'

Comparing the calculated parameters of the steady flow with those of the unsteady flow, we find that their values are quite close and they also correspond with the actual situation in the pumping process. Therefore, the calculated parameters are adopted.

Besides, from the chart of the actually measured waterlevel isolines, we know that the natural hydraulic gradient of groundwater, I = 1 :7,000, is small on the test site. When K = 65.04 m/day, the natural flow velocity of the groundwater of this site is:

V= Kl = 3.39 m/yr

This value shows that the groundwater flow is very slow. Hence, in regions of slowly flowing pressure water where the coefficient values of water transmissibility (T), of pressure transmissibility (a), and of permeability (K) in the aquifer are high, the well-drainage and well-irrigation system can be used effectively.

Multiple-well Test for Drainage

Drainage tests on some siphon artesian wells of different diameters and different fixed discharges were made, and the test results are given in table 8.3.

Calculation of Parameters

The Drainability of Wells and the Calculation of Parameters

Fig. 8.6. Q-S Curve of Injecting Water into Test Well No. 3

We carried out several well-drainage tests in the No.3 well. Totalling 17.5 hr. the total drainage volume was 2,052.88 m whereas the average drainage volume was 117.31 m/h. The average water-level raised was 1.259 m in the well. At one time, when the drainage duration was the longest, the average discharge was 144.02 m/h with a highest raised water level of 2.375 m. After the well drainage had been stopped, the turning points on the temporal curve showing the restoration of water levels were very clear, and water returned rapidly to a level near the original one. The curve Q = f(S) during the process of injecting water approximates the parabola type (fig. 8.6). Owing to the blockage of the lower part of the well and the initial impact of siphoning, the points (Q. S) did not fall completely on the curve.

The various hydrogeologic parameters under the condition of welldrainage were calculated using the same method as for water pumping. The calculated results, using well G. as an example, are shown in table 8.4.

Comparing the hydrogeologic parameters under drainage with those of pumping, we find that they are similar. Mainly as a result of blockage in the well, the values of T. a, and K of drainage are smaller than those of pumping. The elastic waterstorage coefficient is a little higher than the elastic water-release coefficient, a function, probably, of the general decline of the regional pressure head in recent years.

A good indicator of the drainage capacity of the tubular well is the quantity of specific drainage, which refers to the quantity of water drained from a tubular well where the water-level has risen 1 m. When the ranges of change of the water-levels during injecting and pumping approximate each other (but in opposite directions), the initial quantity of specific drainage is either close to or slightly larger than the initial specific yield of water. But the quantity of specific drainage is generally unstable and declines progressively with time, whereas the latter is stable and has a higher value.

The Decay and Restoration of the Drainage Capacity of the Tubular Well

The curves of Q-T and S-T in the process of tubular-well drainage do not show the same rapid appearance of the steady states that are present during pumping (figs. 8.4 and 8.7). Quasi-steady states, however, do appear quickly, and are characterized by a fairly rapid initial rise in the water-level followed by a slow and uniform rise. The drainage quantity decreases gradually towards the decaying state. Drainage tests show that the more turbid the water the faster will be the decay. This indicates that the main cause of decay is the blockage of the filter in the well. Reversed pumping after drainage. however, can make the quantity of specific drainage return to normal. The highest turbidity in the reversely pumped water appears at two minutes after reversed pumping. Apparently the blockage takes place mainly around the filter and can be effectively removed by reversed pumping.

Drainage tests of different durations show that the drainage time should not be too long. Generally it should be about eight hours each time, followed by twenty minutes of reversed pumping. The climatic characteristics of the Sanjiang Plain dictate that drainage and irrigation should alternate in different seasons. Usually the pumping time is longer than the well-drainage time, and this is favourable for well maintenance.

The Influence of Well-drainage on the Quality of the Groundwater

The groundwater of the Sanjiang Plain is mainly soft freshwater with HCO3-Ca-Na and a low level of mineralization. The swampwater belongs mainly to the bicarbonate type, but in some places is of the HCO3-SO4-Ca-Mg or HCO3-CI-Ca Na type. The level of mineralization in the swamp-water is less than 0.1 9/l, which is lower than that of the local groundwater.

Fig. 8.7. Q-Tand S-TCurves of Injecting Water, Well No. 3

The water quality is quite good. Except for the relatively high ion content, the contents of NO3, NO2, Cl- and of heavy metals are not beyond the standards of drinking water. The water flowing back to the farmland in the test site area belongs to the HCO3-KNa type, with a level of mineralization of only 0.032 9/l. This shows that the water drained by wells from swamps and from the cultivated land will not cause salt accumulation; on the contrary, it will dilute the groundwater. In the western part of the plain, swamp-water and groundwater are linked. After rain, the water on the cultivated land infiltrates very rapidly, diluting the groundwater. Therefore, drinking water is normally obtained from shallow wells.

The O2 consumption rate in the groundwater of the test site is 1.22-3.69 mg/l, which is modest, whereas that of the swampwater is mostly 3-9.4 mg/l (18.46 mg/l in some cases). In Heilongjiang Province the standard of O2 consumption in drinking water is required to be 10 mg/l. In some countries, it is 10-40 mg/l. By comparison, the O2 consumption of the swamp-water in the Sanjiang Plain is not high.

The organic content of the groundwater is 2.97-7.3 mg/l (11.95 mg/l in some cases). The organic acid content is 0.07-1.95 mg/l whereas that in the swamp-water is generally 1 mg/l. The water flowing back into the well from the cultivated land contains more organic matter, thus it should be treated before well drainage.

The Efficiency and Benefits of Well Drainage and Well Irrigation

The area of swamps drained by a single well with a diameter of 400 mm was calculated by:

where F= the area (in ha) of swamps drained by a single well per day,

Q = the average discharge quantity of a single well: 144.02 m/day,
t= the effective time of well drainage: 23 hours/day,
Ho = the average depth of swamp-water: 0.1 m;
Hp = the thickness of the grassroots layer: 0.2 m,
K = the porosity of the roots layer: 80 per cent,
h = the coefficient of water discharged by channels: 0.8.

The computed result is F = 1.59 ha, namely, a single well can drain 47.78 ha/month of marshland.

According to a report on the hydrology of the Sanjiang Plain prepared by the Office of Comprehensive Planning of the Sanjiang Plain, the depth of storm runoff caused by heavy rains which generally appear once every five years is F 20% = 25 mm. After a day of such heavy rain, two days are needed for drainage. Under such a condition, the area of cultivated land that can be drained by a single well is calculated as follows:

On the other hand, the area irrigated by a single well is computed as follows:

where Q= the designed water yield of a single well:
169.96 m/hour,
to = the effective time of water pumping: 23 hours/day.
t = the period of rotating irrigation: 10 days,
Qi = the fixed amount of water for irrigation: 600 m/ha,
K = the coefficient of atomization: 5 per cent,
h = the coefficient of water discharged by channels: 80 percept.

The computed result is that a single well can irrigate an area of 49.5 ha in a period of rotating irrigation (10 days).

Because the well-drainage and well-irrigation method can play the dual role of relieving drought and waterlogging, it is obviously beneficial to the increase of agricultural output and to achieving stable and high yields. In 1976 drought was serious on the second experimental site during the jointing stage of wheat. Spray irrigation with pumped water was used in time. But at the time of the wheat harvest, rainfall increased suddenly. On 6 August, a major storm brought 115 mm of rain, and as a result the water content in the top 5 cm of the soil reached 61.3 per cent, and all ditches and low-lying areas were filled with water. Using a well with a diameter of 300 mm for drainage,1,121.7 m of water were drained in 28 hours, and the water content in the soil dropped to 42 per cent. Waterlogging was thus controlled in good time. Another rain of 31.5 mm fell on 19 August, but waterlogging was also controlled through eight hours of well drainage, and mechanical harvesters could be used the next day. In comparison, because of water accumulation in some higher fields in neighbouring areas, harvesting had to be done by hand, which delayed it by more than 20 days. In that year, the average yield of wheat in this experimental area was 32.9 per cent higher than that of the control fields, and the per-unit increase in soybean output was 18.6 per cent. When compared with fields reclaimed in drought years and abandoned to renewed swampiness in waterlogged years, or those on which the use of machinery for harvest was impossible, the benefits of the well-drainage and wellirrigation method are even more apparent.

Groundwater and Well Drainage

Our investigation and calculations show that the velocity of groundwater in the test area is rather slow, the natural flow velocity being 3.39 m/yr. In addition, the water drained back into the test well did not mix readily with the groundwater, as indicated by temperature differences in the water. in August 1976 a total of 6,233.77 m of water was drained into a well in the second test area. The temperature of the drained water was 1020C. After 10 months, on 22 June 1977, the temperature of the water pumped out from the same well was 11 C, which was 7C higher than that of other wells in the same area. This indicates that the water drained into the well did not immediately flow away as a part of the groundwater but remained near the well for a long period.

Through our tests with the multiple-well system we know that the coefficients of water transmissibility, of pressure transmissibility, and of permeability in the aquifers are all very high. The cause of the stagnation of the groundwater is the slight hydraulic gradient. As soon as injection begins, a cone is formed under the well which leads to a rise in the pressure head near the well and to an increase in the local hydraulic gradient. Thus, in an aquifer with a high coefficient of transmissibility, the injected water quickly squeezes away the water around the well, by means of particle transmission. At the same time, because the coefficient of pressure transmissibility is very high, the pressure head raised by injection exerts pressure in all directions in an attempt to establish a new and balanced pressure head. As soon as injection is stopped the pressure head is quickly restored.

That behaviour of water is also found in drainage tests by injecting water into a group of wells. However, draining water into many wells at the same time involves a large amount of water the subsurface condition of which depends much on the underground storage capacity of the location. Based on the theory of elasticity and contraction for aquifer with pore-pressure, a change in the pressure head will lead to an elastic release or storage of water. When pumping water out of the well a cone of depression is formed around the well, the pressure head is lowered, and the aquifer releases water elastically. On the other hand with draining water into the well an anticone will be formed around the well, the pressure head is increased, and the aquifer stores water elastically. This means that an aquifer with pore-pressure can regulate groundwater storage the capacity of which is affected in part by the level of the pressure head.

According to the theory of dynamic balance of groundwater, groundwater recharge, runoff, and discharge are always in a state of dynamic equilibrium. Data collected by Geological Team 914 in the Fongfu region show that the region's dynamic water storage is 2.15 million m/day. The amount of natural runoff in the flood season of July, August and September is nearly 2 x 108 m. If well drainage is carried out on a large scale in the flood season, which will lead to an increase in the regional pressure head, the amount of natural recharge will then decrease while natural discharge increases greatly in order to establish a new dynamic equilibrium. This would make the amount of runoff in the flood season in this region larger than 2 x 108 m. The new equilibrium could furnish a new dynamic storage capacity for large-scale well drainage.

Drought in spring and waterlogging in autumn dictate the seasons for well drainage and well irrigation respectively. The quantity of water pumped out in spring creates storage space for groundwater to be injected back from the surface in autumn. In the central Sanjiang Plain, where there are 792,000 ha suitable for well drainage and well irrigation, the amount of water needed for two alternate irrigation periods is 1,500 m/ha. In the same area the total water yield per year is 12 x 108 m, which is the equivalent of 150 mm of runoff depth. However, for many years the actual average runoff depth of the area has been only 100 mm. The depth of the maximum rainstorm runoff which occurs once in a decade is only 33 mm. Thus it is entirely possible to regulate the storage capacity by well irrigation and to discharge storm runoff by well drainage.

Through several years of tests and study, we found four major hydrogeological conditions in the Sanjiang Plain that are favourable for well drainage and well irrigation. These are: la) where the aquifer is rich in water, ideal in thickness, and has good water transmissibility; (b) where the depth of the stable water level is moderate; (c) where the conditions of recharge and discharge of groundwater are good; and Id) where the quality of groundwater is suitable for irrigation. The amount of Fe and Mn and the level of pollution in the surface water are low enough that the water does not require treatment before injection.

The Sanjiang Plain has more than 2 million ha of land suitable for well drainage and well irrigation. In most places wells could be used as the primary means to drain wetlands for agricultural development. In some places the well method could serve as a supplementary measure used together with drainage ditches. The effectiveness of the method is readily apparent, and it should be introduced widely.


1 Compilation Groups (eds ), Manual of the Hydrogeology of Water Supply. vol. II, 1977. Geology Press, Beijing (in Chinese.)


Contents - Previous - Next