CHAPTER 5 GROUNDWATER RECHARGE ESTIMATES BY DIFFERENT METHODS

Transcription

1 Groundwater Recharge Estimates for Study Villages 225 CHAPTER 5 GROUNDWATER RECHARGE ESTIMATES BY DIFFERENT METHODS The Saurashtra groundwater recharging movement has been examined from two points of view: one, whether there has been significant enhanced recharge that could provide enhanced agricultural returns to the farmer; two, whether the recharge could be quantified at the village level so as to connect with the agricultural returns. This chapter attempts to address these two aspects. How the enhanced recharge could have contributed to enhanced agricultural returns is analyzed in the following chapter. This Chapter is divided into three sections: section 1 deals with computation of recharge values by three methods, namely, WL & SY Method, the Regression Method and the CRU Method. The theoretical aspects of these methods have been discussed in the preceding Chapter 4. Section 2 examines the Uncertainty in Recharge Estimates by different methods so as to be aware of not only the relative merits and demerits of recharge estimations but also the limitations. Section 3 compares the recharge values obtained from the three methods and draws conclusions on the usefulness of employing more than one method. SECTION 1 ESTIMATION OF GROUNDWATER RECHARGE FOR STUDY VILLAGES [1] Water Level & Specific Yield Method In India, as per the GWRE 2002 guidelines, the CGWB adopts (and recommends) use of the WL & SY method and the Regression Method (CGWB, 2004:27) for computing

2 Groundwater Recharge Estimates for Study Villages 226 water balance and to notify the stage of groundwater development 135 for the purposes of monitoring. However, it also emphasizes cross checking of results between these two methods and stipulates that the variation lie within 20% only (CGWB 2004:33). The guidelines also strongly recommend budgeting for environmental flows and allocations for drinking and industrial purposes 136. The water balance estimate is done by the CGWB at the level of taluka; a taluka comprises many villages and an average figure for such a large geographical area with diverse geological and meteorological conditions is often misleading. The study district Rajkot comprises 14 talukas, 846 villages, and 10 towns and cities 137. The level of groundwater development for Gujarat State was estimated between 65-85% during (GoG, 1997) while it was estimated at 76.47% for the year 2004 (CGWB, 2004). Current indications are that the transition is from semi critical to critical stage (Safe: <70% stage of groundwater development; Semi critical: 70-90%; Critical: >90%; Overexploited: >100% (CGWB & GoG, 2005). Table 5.1 gives the stage of groundwater development for the study talukas of Rajkot district. Table 5.1: Groundwater development in study talukas District Taluka Groundwater development (as of 31 st March 2004) (%) Rajkot Gondal Rajkot Jamkandorna Rajkot Wankaner Rajkot Morbi Source: GoG, The following Table 5.2 gives the detailed steps in this process of computation of recharge for one village Ambaredi. A similar procedure is adopted for all the 6 study villages. The values of recharge obtained by all the three methods are presented in Table Measured as the ratio of gross groundwater draft for all uses to the net available groundwater availability expressed in percentage (CGWB, 1997). 136 These are expected to take care of factors other than the water level and specific yield in the WL & SY equation accessed Feb 10, The reference year 2004 is used as the study year for primary data collection was

3 Groundwater Recharge Estimates for Study Villages 227 Table 5.2: Computation of Groundwater Balance and Stage of GW development for Ambaredi village (Mudrakartha, 2008) WL & SY Regression Method 1 Gross GW Recharge For Environmental Available GW recharge/year (1-2) GW Draft for irrigation from primary data Domestic and Industrial draft (15% of 1) Gross GW Draft for all uses (4+5) Regeneration of (4) Net GW balance (3-6+7) Level (stage) of GW Development (%) [Note : All figures in Million Cubic meters except for sl.no.9; : Minor mismatch in totals due to rounding off may exist] The average annual water-level fluctuation (denoted by the difference between the maximum and the minimum water levels in the wells) for Ambaredi is estimated as 8.1 m for the reference year from June 2003 to May These maximum and minimum water levels in arid and semi arid regions, where unimodal rainfall pattern exists, are obtained generally during peak monsoon and subsequent pre monsoon periods. The water levels actually represent net values as there is extraction as well as indirect or localised recharge possible even after the monsoon duration. Thus, the groundwater recharge computed would almost always be underestimated. Before proceeding for computation, it is useful to understand the terms used. (i) Gross groundwater recharge In general, it is assumed (although not quantitatively measured), that 70% of the natural recharge to groundwater can be extracted (Athavale, 2003). The CGWB guidelines for categorization of groundwater assessment units for water balance also consider up to 70% groundwater development as safe (GoG, 1998; CGWB, 2004:18; Planning Commission, 2007). Table 5.2 gives the detailed steps for computing the groundwater balance. The specific yield of basalt, the predominant rock type in the study areas, ranges from 1 to 3 (Sinha & Sharma, 1988). Using equation the WL & SY equation (1) [NR = Area (sq. km) x

4 Groundwater Recharge Estimates for Study Villages 228 average of difference of maximum and minimum water levels from wells (m) x specific yield (%) of the geological formation], the volume of rainwater that has recharged into the ground (NR) is computed as MCM and 6.67 MCM for specific yields 1 and 3 respectively for basalt which is a very wide range. Which value of the specific yield would represent the study village is discussed in section 3? For our present purpose, we assume that the specific yield 1 may be appropriate, also being on the lowest side, in order to reduce uncertainty due to reasons explained in section 2 of this Chapter. Using the second method of Regression equation, [natural recharge NR (or RE) = (Rainfall in mm)-62], the volume of water that has recharged into the ground is computed as 1.83 MCM. In terms of recharge, the two methods yielded values given in columns B, C and F of Table (ii) Allocation for environmental flows GoG guidelines of 1997 & 2002 (GoG, 1998; CGWB, 2004) provide for 5% of gross groundwater recharge for surface flows in river bodies for inter-basin transfers. Concerns during the eighties and the nineties were that many important basins in India were either closing 140 or closed 141. Some studies have also expressed concern about many sub basins also being in the process of closing or closed due to the large number of water harvesting activities promoted by various agencies. These activities intercept not only the surface flow to downstream areas but also impact groundwater recharge adversely. The GWRE guidelines (2002) provide for a 5% allocation of gross groundwater recharge as environmental flows to check, and pre-empt, partial or complete closure of basins. As can be seen from Table 5.2, the computations here consider this allocation. (iii) Groundwater draft Groundwater draft represents the total volume of water pumped out from all the well structures in a particular year. The draft can be computed based on the well inventory data comprising, inter alia, the number, and type, of pumpsets and their horsepower, their running hours and discharges, for all wells in a given village. Table 5.2 shows the total 140 Closing basins are those that have inter basin flow during wet season, but no usable flow during dry season. Even the Indus, the Ganges and the Yellow River are closing by this definition (Seckler et al. 2003). 141 Closed basins are those that do not have any usable flows during any part of the year, not even during wet season (Seckler et al. 2003).

5 Groundwater Recharge Estimates for Study Villages 229 annual groundwater draft for Ambaredi village computed as MCM for the year (iv) Allocation for Drinking Water and Industrial needs GoG guidelines of 1997 & 2002 (GoG, 1998; CGWB, 2004) also provide for 15% of the annual gross groundwater recharge towards drinking water and industrial needs. Under these guidelines, the competent authority 142 could regulate extraction of groundwater in times of water scarcity, thus meeting essential drinking water needs. Where industries do not exist, which is usually the case in rural areas, the water is supposed to be available for environmental purposes. In practice, it may also be overextracted. (v) Regeneration Recharge or Return Flow The irrigation provided to crops is not fully utilised by the crop. There is return flow 143 into the water bearing formations or aquifers. It is estimated that one-third of the total water used for irrigation percolates and adds to the groundwater reserve (Athavale, 2003). For the purpose of computations for the study villages, thirty percent of return flow is considered also because of the shallow groundwater level (near surface during monsoon) with a high at 20 m for Ambaredi and 13 m for Jalsikka cluster of villages. The irrigation practice is through field channels covering the whole farm; the channels are of course not too wide to be classified as suitable for flood irrigation. [2] The Regression Method The regression equation for basalt is Recharge RE=0.174 (Rainfall in mm)-62 (Athavale, 2003). Parameters other than rainfall are subsumed to be taken account of. On the face of it, mathematically, rainfall is the only input given. For the study villages, average rainfall of 740 mm for Rajkot district is considered for the year 2003 as obtained from official sources for computations. This gives a uniform value of recharge of mm for all the study villages. 142 The competent legal authority is the Central Ground Water Authority, which functions through its Regional Offices; often, the district administration is notified as local authority. 143 Seepage and percolation losses are about 60 percent of the applied water in a canal command area (Reddi and Reddy 2006).

6 Groundwater Recharge Estimates for Study Villages 230 [3] CRU-NUT_MONTH Method Chapter 4 has described the Climatic Research Unit (CRU), University of East Anglia proposed method of estimating recharge using climatic variables such as precipitation, temperature, diurnal temperature, vapour pressure, cloud cover, sunshine duration and wet days. The procedure and steps for estimation of recharge have also been described in detail therein. The CRU and NUT_MONTH method is very useful as in the absence of actual hydrological data such as observations of river flow data at a number of points along a river and its tributaries; long period basic meteorological data like rainfall, temperature, humidity, etc., can be used for estimating water potential of a region or a basin and its variation in space and time by using suitable technique (Kulkarni, 2003 as quoted in Ramesh & M. G. Yadava, 2005). The method is also useful to assess groundwater recharge when water level data is not available. The following steps are adopted for running the CRU Model: The location coordinates of the areas for which climate variables are sought to be extracted are identified; the coordinates can be obtained from a toposheet, published by the Survey of India. The coordinate referencing in CRU program is made in the form of degrees. Therefore, the coordinates are converted into degrees for easier working as shown in the 5. 3 (last two columns, (5) and (6)) for the study villages. As can be seen from the same Table 5.3, the study villages are falling in two distinct nodes in the CRU global database. Ambaredi village falls in first node (hereafter referred to as Ambaredi cluster), and Jalsikka, Vithalpar, Haripar, Kerala and Bella (hereafter referred to as Jalsikka cluster), fall in second node (column 2).

7 Groundwater Recharge Estimates for Study Villages 231 Table 5.3: Study Villages and Coordinates S. Village Latitude Longitude Cell Reference node no. (in deg., (in deg., coordinates on CRU global min., sec) min., sec) (in degrees) database (1) (2) (3) (4) (5) (6) a. Ambaredi 22_21-71_ E N E, N E, N b. Jalsikka 23_22-71_ E N 71.00E, N E, N c. Vithalpar 23_22-71_ E N E, N E, N d. Haripar-Kerala- Bella 23_22-71_ E N E, N E, N In the CRU input file, the location coordinates along with village name are provided as shown in the (CRU_input file), and the years for which the climate data is desired. Running the program CRU_TS21_READ_METEO by double clicking would use the CRU_input file and generate an output file that contains data organised in four files corresponding to four sub-cells. Consider an area represented by latitude and longitude of one degree by one degree; the CRU program divides each such area into a cell of 0.5 deg. by 0.5 deg. Thus, we have four sub-cells in one degree by one degree. The data is organised around the mid-point of each sub-cell; the outputs from CRU are also generated around the mid-values of these four sub cells. Thereafter, the sub-cell that represents the study area needs to be identified since the resolution of the CRU program is 0.5 deg by 0.5 deg. The last column of the Table 5.3 represents the sub-cells for the study villages in Rajkot district. As already mentioned elsewhere, the output file of CRU contains month-wise precipitation and number of stations corresponding to the years specified, for the four sub-cells. The data of the subcell that comprises the location of the study area becomes an input file for the NUT_MONTH program that computes precipitation recharge.

8 Groundwater Recharge Estimates for Study Villages 232 LONG TERM RAINFALL-RECHARGE ANALYSIS OF STUDY VILLAGES BASED ON CRU DATA In a Research Report (No: 2/2006) published by the National Climate Centre, P. Guhathakurta and M. Rajeevan (2006) analysed the Trends in the rainfall pattern over India. They cite several studies that revealed no particular trend, rather it was random behaviour of the Indian monsoonal rainfall. But on spatial scale, trends were noticed. On an all India level, the months of June, July, August and September contributed 13.8, 24.2, 21.2, and 14.2% to the total rainfall respectively. The post and pre monsoon rainfall contributed 11% each. Although broadly the monsoon rainfall is categorised as a systematic event occurring every year, it shows considerable variation during individual years. The important aspects of the variations are (Guhathakurta, P & M. Rajeevan, 2006): The timing of the onset or the commencement of the rainy season; The pattern of distribution of rainfall including the timing; The timing of withdrawal of the monsoon from the different parts of the country, and The total amount of rainfall of the season. For a farmer, both the total rainfall and its distribution during the rainy season are important. The common challenges faced by farmers include delayed or early commencement of the monsoon rains, long breaks comprising no rains, and early withdrawal. High intensity spells or excessive spells also result in flooding and water logging, and consequently, loss of crops, partially or fully. As discussed earlier, the pattern of rainfall also determines the rate of recharge. Based on the analysis of long term rainfall data for Gujarat for the years 1901 to 2002, Patel, K. I et al. (2004) indicated that the consecutive years of receiving negligible to below normal rainfall never exceeded more than three years; whereas the consecutive

9 Groundwater Recharge Estimates for Study Villages 233 years of having received near normal to above normal ranged from one to five years in Saurashtra and Kutchh region144. Further, Indian monsoonal rainfall has a high coefficient of variation, CV, (standard deviation expressed as percentage of the mean) exceeds 30% over large areas of the country and is over 40-50% in parts of Saurashtra, Kutch, and Rajasthan. In some places, the variability is as high as 100% implying that these places are particularly liable to very heavy rainfall in some years and very scanty rainfall in others (Jagannathan & Bhalme, 1973). For example, the lowest and highest rainfall recorded during the period respectively were 144 mm and 1361 mm for Ambaredi, and 120 mm and 1328 mm for Jalsikka clusters. In addition, there is high yearly fluctuation as can be seen from Figures 5.2 and 5.8. Therefore, these variations should be kept in mind while drawing conclusions, acknowledging the inaccuracy and uncertainty elements that will be introduced, in particular, during computation of recharge values. This section examines the long term trend analysis of rainfall and recharge, and then the rainfall-recharge relationship for both clusters of Ambaredi and Jalsikka villages. Further, the factors Actual Evapo-transpiration (AET) and Potential Evapo-transpiration (PET) would also be considered and their influence on the natural recharge process will also be examined. As can be seen from Table 5.3, the study villages are falling in two distinct nodes in the CRU global database. Ambaredi village falls in first node, and Jalsikka, Vithalpar, Haripar, Kerala and Bella fall in second node. The rainfall data for Ambaredi and Jalsikka clusters for 102 years-from 1901 to 2002 is extracted from the public access domain using the CRU_TS-READ_METEO program as explained earlier. The rainfall and recharge analysis are carried out for both the clusters as described in the following section. 144 For other regions in Gujarat: one to seven years in North and Middle Gujarat region, and one of seventeen years in South Gujarat region. The Kutchh, Surendranagar, Banaskantha, Patan, Mehsana, Jamnagar, Kheda, Anand, Rajkot and Bhavnagar districts experienced drought conditions two to three times during the period 1991 to 2002.

10 Groundwater Recharge Estimates for Study Villages 234 The data for Ambaredi shows that the lowest rainfall was 144 mm corresponding to the year 1987 and a high 1361 mm in the year The average long term rainfall based on the rainfall data for Ambaredi works out to mm. Table 5.4 shows that once in 12.8 years, we have a situation of rainfall less than 300 mm. On an average, Ambaredi receives rainfall between mm once in 2.7 years, and between mm, once in two years. Put differently, once in 2.3 years, there is a probability of rainfall between 301 and 1000 mm (see also Figure 5.3). Table 5.4: Rainfall data for for Ambaredi cluster No. of Years Frequency (once in.. Rainfall mm years in % years) Total Figure 5.1 is scatter diagram between rainfall and recharge for Ambaredi for the long term data of 102 years. The rainfall on the x-axis is sorted in ascending order and recharge values plotted indicates that there is a broad correlation in the trend between rainfall and recharge. A detailed analysis shows that recharge is generated only under certain conditions of rainfall; soil constants and climate parameters also play a role. This section examines these aspects for the study villages.

11 Groundwater Recharge Estimates for Study Villages Rainfall-Recharge for Ambaredi Figure 5.1: Scatter diagram between rainfall and recharge for Ambaredi Figure 5.1 shows that for rainfall of up to around mm, the recharge generated is almost negligible-in other words, close to zero. The number of such zero-recharge years is 25 out of 102 implying a frequency of once in 4.1 years as seen from Table 5.5. Further, the figure also shows that between mm, there are many years with recharge around 60mm, although interspersed with zero recharge. Beyond a rainfall of 500 mm, the recharge tends to become more certain. For a rainfall range of mm, which is quite populous, the recharge is around 100 mm-something very significant for agriculture in Ambaredi, given its semi arid climatic conditions. Beyond 640 mm, the correlation between rainfall and recharge becomes much more positive, with no zero recharge years; the recharge on the contrary tends to become significant.

12 Groundwater Recharge Estimates for Study Villages Rainfall-Recharge for Ambaredi Rainfall, mm Recharge, mm Figure 5.2: Rainfall-Recharge relation for Ambaredi 60.0 Rainfall versus prob of RF occurrence, frequency- Ambaredi % years frequency: once in years Rainfall in mm Figure 5.3: Rainfall versus probability of RF occurrence and frequency years.

13 Groundwater Recharge Estimates for Study Villages 237 Table 5.5: Recharge and frequency of recharge occurrence for Ambaredi Recharge mm No. of years % Frequency (1 in.. years) >100mm Let us analyse from the point of view of the recharge pattern in Ambaredi. As can be seen from Table 5.5, zero or negligible recharge occurs in 25 years out of 102 years: this means that once in 4.1 years, there will be at least one zero or negligible recharge year (corroborates with the point that the frequency of rainfall is also zero in one out of four years). The zero or negligible recharge year also corroborates with water scarcity or drought year. The highest recharge of 609 mm has occurred in the year 1959 for the highest rainfall of About 60 mm recharge is possible once in 5.4 years. If we consider recharge between 61 and 300 mm, which is a significant quantum of recharge to occur in semi arid areas, the probability works out to once in two years. Higher recharge is always welcomed. What is the frequency of occurrence of more than 100 mm recharge? Table 5.5 further shows that 47 out of 102 years have generated recharge greater than 100 mm, which translates roughly as once in 2.2 years. Very high recharge of 300 mm and above occurs once in 8 years; although spaced, such wet years tend to build up groundwater storage, as can be seen from the rainfall-recharge relationship (Figure 5.1), in some way compensating for secular declines in the local water levels.

14 Groundwater Recharge Estimates for Study Villages 238 Thus, it may be generally concluded that, for Ambaredi, a rainfall of 500 mm and beyond will generate a recharge of at least 60 mm, and a rainfall beyond 640 mm will generate 100 mm. Since such measures of recharge also imply abundant soil moisture, agriculture is expected to benefit significantly, subject to water-intensive crops not raised majorly. During the other years, for rainfall between mm, some recharge does take place. Similarly, although Figure 5.1 indicates that the possibility of recharge to groundwater in Ambaredi is almost nil below say, mm, in reality one could expect some amount of recharge seen in the form of quick build up of water levels in response to even say a couple of high intense spells when occurring, in particular in hard rock areas such as Ambaredi, where water tables are shallow. This aspect indicates that the values obtained as recharge also have an inherent element of uncertainty, sometimes interpretation errors adding to this uncertainty; this element however seems to be very small and negligible. Similar discussion is valid for Jalsikka cluster as described in the following section. Recharge versus RF probability-ambaredi Recharge in mm Recharge in mm Figure 5.4: Recharge versus Rainfall probability-ambaredi

15 Groundwater Recharge Estimates for Study Villages Ambaredi Potential ET in mm Recharge in mm Actual ET in mm Rainfall in mm Note: Curves PET and AET overlap, shown as the top curve. Figure 5.5: Relationship between rainfall, recharge and PET and AET for Ambaredi The Actual Evapotranspiration (AET) and the Potential Evapotranspiration (PET) are important climatic factors in groundwater recharge. Figure 5.5 indicates that the PET and AET overlap completely because the values are exactly the same every year; the ratio of AET/PET therefore is equal to one. This implies that the field capacity is achieved and recharge taking place on a year to year basis. However, Table 5.5 also shows that there is nil recharge once in four years. Which implies that there should be at least one water deficit year out of 4.1 years. However, when we analyse the rainfall data from Table 5.4, it is clear that less than 300 mm rainfall occurs once in 12.8 years. Table 5.5 also shows that there is recharge between mm happening once every 2.1 years, more specifically, more than 100 mm once every 2.2 years. So, we have a situation where there is some soil moisture retention taking place even during low rainfall years (say 300 or 400 mm), and no recharge being generated. However, as discussed in the foregoing, the rainfall mm range is quite populous (Figure 5.1), and contributing to generation of recharge. All these factors point to the soil moisture availability of some degree at the end of the year which is most probably carried forward to the next hydrological year. Whether this soil moisture contributes to the recharge directly is not known, but it does hasten the recharge by way of achieving quicker saturation of soil (that is, field capacity that includes root zone) during the succeeding rainfall events. The available soil moisture

16 Groundwater Recharge Estimates for Study Villages 240 may not be of any help to the farmer in making sowing decisions as he would not have methods of knowing or estimating soil moisture content in the root zone. However, when the sowing decision is taken at the advent of timely first rains, this soil moisture would help by prolonging the wilting point of the crop. Alternatively, it may be also be concluded that though the soil moisture condition in general is good in Ambaredi, and recharge does happen during 3 out of 4 years to support crops, there is some inadequacy in the soil moisture balance method itself to reflect accurately the ET variations on a year to year basis. The AET and PET values are estimated by the Thornthwaite-Mather method in the NUT_MONTH programme as the values were broadly matching with the values already referred to in literature. Also because the normal method of estimating the ET basing on assumptions of crop type, area, soil conditions etc. which are generally inadequate introduce inaccuracy in the recharge estimation. In short, if we consider a four year cycle, two years could be with a rainfall of mm, one between mm and one less than 300 mm. In terms of recharge, one could be a year with zero recharge, two with recharge between 60 and 300 mm, and one could be with minimal or recharge less than 60 mm. This is a broad trend in terms of rainfall and recharge pattern based on long term analysis. It is interesting to note that the farmers of Ambaredi during focus group discussions have found that their recharge activities have helped them take at least two crops every yearone among them being the 6-month cotton crop. Rainfall and Recharge Analysis for Jalsikka Cluster The rainfall data for Jalsikka cluster of villages (Jalsikka, Vithalpar, Haripar, Kerala and Bella) is sourced from the public access domain The long term average computed from the long term data works out to mm. The data reveals that Jalsikka cluster registered a low rainfall of 120 mm for the year 1987 to a high of 1328 in Analysis of rainfall shown in Table 5.6 indicates that once in 6.8 years, Jalsikka cluster

17 Groundwater Recharge Estimates for Study Villages 241 receives rainfall less than 300 mm. The rainfall incident is between mm once in 2.6 years on an average, and between mm once in 2.5 years. Put differently, once in 2.5 years approximately, there is a probability of rainfall between 301 and 1000 mm. Table 5.6: Analysis for Rainfall trend for Jalsikka Cluster Rainfall mm No. of years Years in % Frequency (1 in.. years) Rainfall vs prob of RF and frequency-jalsikka % probability frequency: once in..years Rainfall in mm for Jalsikka Figure 5.6: Rainfall vs. rainfall frequency and probability of occurrence for Jalsikka When we consider rainfall and recharge pattern of Jalsikka cluster as shown in Figure 5.7, it is difficult to draw a strictly linear correlation between rainfall and recharge on a year to year basis, just like in the case of Ambaredi. However, there appears to be a broad correlation as can be seen from Figure 5.8 for the given soil and climatic conditions of Jalsikka.

18 Recharge in mm Recharge in mm Groundwater Recharge Estimates for Study Villages Rainfall in mm for Jalsikka Fig 5.7: Rainfall-Recharge relation for Jalsikka Rainfall in mm for Jalsikka Rainfall in mm Recharge mm Figure 5.8: Rainfall-Recharge correlation for Jalsikka Further analysis of the rainfall in relation with recharge from Figure 5.7 (Rainfallrecharge correlation) would reveal that a rainfall up to 560 mm approximately, has produced negligible recharge. The non-zero recharge years for Jalsikka cluster below 560 mm of rainfall are very few, unlike in the case of Ambaredi cluster. Put differently, the number of zero or negligible recharge years for Jalsikka cluster is 40 out of 102, which works out to once in 2.6 years (Table 5.7) as compared to Ambaredi which is once in 4.1 years. Figure 5.8 shows that rainfall above 560 mm and below 620 mm has produced an

19 Groundwater Recharge Estimates for Study Villages 243 annual recharge of around 60 mm. Between 620 mm and 750 mm, the recharge is generally around mm. For a rainfall of above 750 mm, the recharge generated is above 100 mm. If we look at recharge quantity and frequency, less than 60 mm recharge is possible once in 7.8 years, excluding the zero recharge years, as can be seen from Table 5.7. If we consider recharge between 60 and 300 mm, which is a very significant quantum of recharge to occur in semi arid areas, the probability works out to once in 2.6 years. Further, as can be seen from Appendix 3, the highest recharge of 593 mm occurred during the year 1956 which is also the year of highest rainfall of 1328 mm. What is the frequency of occurrence of more than 100 mm recharge? Analysis indicates that 37 out of 102 years have generated recharge greater than 100 mm, which transforms into a frequency of roughly once in 2.8 years; between mm recharge occurs once in 8.5 years. The graph (Figure 5.8) indicates that approximately mm of recharge is produced for rainfall range of mm. Higher recharge of 300 mm and above occur infrequently; however these wet years tend to build up groundwater storage, compensating for long term declines in water levels. Table 5.7: Recharge in Jalsikka No. of Frequency (once in.. Recharge mm years % years) >100 mm

20 Groundwater Recharge Estimates for Study Villages 244 Groundwater Recharge and Crops Groundwater recharge is a function of the soil properties, temperature and evapotranspiration (that includes vegetative cover/landuse-in the case of Jalsikka, the presence of crop, mostly) some of which keep altering on a year to year basis. For agriculture crops to be raised, a combination of rainfall and recharge (in the form of water in the wells) are important. The rainfall helps in land preparation and timely sowing while the recharged water from wells supports supplementary irrigation during rabi and during long intervals of rainfall, or low rainfall during kharif. The common crops raised in the study villages are cotton, groundnut, and winter wheat. Cotton is taken as a 6-month crop-sown in kharif and goes upto rabi. It is interesting to note that the recharge (when occurring as discussed in the previous section) for a rainfall window is always a percentage lower for Jalsikka compared to Ambaredi. In Jalsikka, Vithalpar, Haripar, Kerala and Bella cluster of villages, the wells are of around 13 m depth; lithomarge of 1-2 m thickness occurs anywhere between metres. The top soil is clay. Conditions here seem to facilitate recharge only during longduration, high intensity spells, and aided by soil and moisture conservation structures, both in-land and across the streams (constructed as part of the watershed programme) during which the soil reaches its field capacity and soil moisture adds to the shallow water table. In case of smaller duration, high intensity spells, there is soil moisture that is added. In other rainfall conditions such as when it is of very high intensity, more runoff is indicated perhaps indicated by the large number of zero recharge years. The addition of soil moisture during smaller duration, high intensity spells increases the available soil moisture, discussed and corroborated in later sections. The wet years, though interspersed at longer intervals of time and whenever occurring, contribute to groundwater recharge by raising water levels. This can be seen in the fact that recharge greater than 100 mm is possible once in 2.8 years for Jalsikka cluster, which is 2.2 years for Ambaredi.

21 mm Groundwater Recharge Estimates for Study Villages 245 The soil moisture balance is a critical factor in recharge generation. The occurrence of recharge is also dependent upon the carried forward soil moisture balance. Farmers in Haripar, Kerala and Bella take one crop mostly, the 6-month cotton crop and groundnut. Half of the farmer community goes in for low water requiring crop as second crop. Further local situations also could be influencing the soil moisture balance, and the water levels in the wells, through stream-aquifer connectivity such as in the case of Jalsikka and Vithalpar, mainly due to availability of water in the river for longer time duration. What about the evapotranspiration in Jalsikka? Figure 5.9 shows the potential and actual evapotranspiration graphs drawn from the values obtained from the NUT_MONTH programme for Jalsikka Jalsikka Precipitation in mm Potential ET in mm Actual ET in mm Recharge in mm Year Note: Potential and Actual ET values are same and hence only one curve is seen. Figure 5.9: Potential and Actual Evapotranspiration trend for Jalsikka It can be seen from Figure 5.9 that the AET and PET values for Jalsikka are exactly same-identical every year although year to year variation exists. Because of this, the two curves coincide perfectly. The year to year variation, however, is not very large. Same values of PET and AET, or the ratio of AET/PET equal to one, indicate that the soil moisture supply is sufficient (Tallaksen et al. 2004). This is also corroborated during focus group discussions with Jalsikka farmers when they said that they are able to take only one or two crops. It is common for Jalsikka group of farmers to go in for cotton

22 Groundwater Recharge Estimates for Study Villages 246 which is a six months crop; in effect this is like taking two crops. Though there is some water left in the wells, farmers feel that it is insufficient for irrigating a whole water intensive third crop. A small percentage of farmers raise crops such as vegetables and small millets etc. Sensitivity Analysis for soil parameters In situ weathering of pre-existing rocks produces sand, silt and clay in combination that is dependent upon the composition of the parent rock type. The shape, size and roundness of the resultant particles depends upon the degree of weathering and extent of transportation they have been subjected to. Depending upon the combination of the sand, silt and clay, the soil types are categorised as fine sand, sandy loam, silt loam, clay loam and clay. The soil textural classification chart has been developed by USDA which shows clay on one extreme, with 40% of clay particles and 45% sand, and the rest silt; in the intermediate level, loam contains equal composition of sand, silt and clay. There are many types of the clay mineral; depending upon the type of clay mineral, the soil displays characteristics of swelling or shrinking with changes in water content. The other combination would have sand and silt in higher percentages. Loam soils are considered to be most favourable for plant growth as they can hold more water than sand and better aerated than clay; they can be worked easier for land preparation for agriculture. The composition of the soil, in other words, the texture, determines the composition of the other two phases, namely, the soil water and soil air phases as described earlier (Michael, 1983). These aspects have been discussed in section 1 of Chapter 4 on Groundwater Recharge Literature Review. The USDA classification chart (given in standard text books) provides a method of identifying the type of soil depending upon the composition. The water holding capacity of the soils is as follows (Michael, 1983):

23 Groundwater Recharge Estimates for Study Villages 247 Table 5.8: Water holding capacity and soil type Soil type % moisture, based on dry weight of soil Depth of available water per unit of soil cm/ m depth of soil Field Capacity Permanent (FC) Wilting point (WP) Fine sand Sandy loam Silt loam Clay loam Clay The objective of this section is to understand the influence of the soil parameters and the rooting depth on the available soil moisture and recharge. Tables 5.8 and 5.9 basically give output data from the program NUT_MONTH for various inputs of rooting depth and the soil type described by its clay, sand and silt composition. The input parameters also include number of soil layers and the FC, CP and WP in percentage appropriate for the crops raised in Ambaredi and Jalsikka clusters. The output data includes the FC, CP, WP in mm, the climate data, and the recharge. The available soil moisture can be calculated as the difference between FC and the WP, as shown in the Tables 5.8 and 5.9. Relation between Rooting depth, FC and Recharge In order to study the relationship between rooting depth, FC and recharge, let us consider Table 5.9, and Figure 5.10 generated from it for the Jalsikka cluster. Various rooting depth values (10, 50, 90, 120 cm) were used in the NUT_MONTH programme to compute FC, recharge etc. Table 5.9: Rooting Depth, soil constants, recharge and available soil moisture relationship for Jalsikka for the year 2003 for actual rainfall (740 mm) Roo ting dept h Rai nfal l PET AET Cla y San d Sil t F C C P WP FC CP WP Rech arge Cm % % % % % mm mm mm mm mm mm mm mm mm a b c D e Avl soil moist ure

24 Groundwater Recharge Estimates for Study Villages Rooting depth vs FC andrecharge-jalsikka Actual RF Recharge in mm 57 FC in mm Rooting depth in cm Figure 5.10: Rooting depth, FC and Recharge for Jalsikka cluster Figure 5.10 shows that as the rooting depth increases, recharge decreases. This implies that the thicker the root zone the higher is the demand for soil moisture. Which implies that after saturation of the rooting depth, the soil zone would tend to achieve field capacity under conditions of adequate water supply and result in soil moisture balance and recharge. Recharge would also depend upon the depth to the water table. Here in the case of Jalsikka, the recharge is occurring, which indicates that field capacity is achieved and the water is reaching the groundwater table. Well inventory indicates that the depth to the water level during monsoon in wells is near surface or shallow depending upon the physical elevation of the well. Further, the rooting depth and recharge are inversely related, which also implies that the recharge will depend upon the type of crops grown, because different crops have different rooting depths depending not only on the crop but also on crop variety (Michael, 1983). Put differently, this implies that the field capacity should increase with rooting depth which is corroborated by the rooting depth versus FC curve in the Figure Again, increased rooting depth also implies increased available soil moisture readily accessible for the plant, which is computed as the difference between the field capacity (FC) and the permanent wilting point (WP). This is corroborated by the Figure 5.11 for Jalsikka for the year 2003.

25 Groundwater Recharge Estimates for Study Villages Rooting depth vs Recharge and Soil moisture-jalsikka: 260 Actual RF 2003 Recharge, mm Rooting depth, cm Figure 5.11: Rooting depth versus available moisturefor Jalsikka cluster. The available soil moisture is also calculated for all the situations. Fine-textured soils have a wide-range of water between FC and permanent WP than coarse-textured soils unlike sandy soils which have mostly non-capillary water that tends to release most of it within a narrow range of potential due to predominance of large pores (Michael, 1983). As can be seen from Table 5.8, the range of water available as soil moisture is mm for clay and mm (16-30 cm) per metre depth for clay loam. This availability reduces as the silt and sand composition increase. Fine sand at the end of USDA chart opposite clay has just mm of available soil moisture. Put differently, the Poiseuille s law comes to play where the rate of flow of water (through a pipe) is proportional to the fourth power of the pore size. Assuming saturated conditions, since the pore size increases as the transition happens from sand to loam to clay, the rate of flow in soils of various textures is also more or less in that order (Michael, 1983). In short, soils with fine texture serve not only as good storage zones but also yield larger quantities of water. What is the relationship of WP, recharge and available soil moisture? While these are touched upon in the above discussion for Jalsikka, we will discuss based on the output data obtained from the NUT_MONTH programme for various inputs of wilting point for the study villages. Although the pattern of response is similar for all the study villages, for the purpose of discussion, let us consider the Table 5.9 again for Jalsikka. The Table and the Figure 5.12 shows that as the rooting depth increases, the wilting point increases and consequently, the available soil moisture increases as also the recharge.

26 Groundwater Recharge Estimates for Study Villages Rooting depth vs Wilting Point, Recharge and Soil moisture-jalsikka: Actual RF Rooting depth, cm Recharge, mm Soil moisture, mm Wilting point, mm Figure 5.12: Relation between Rooting Depth, Recharge, soil moisture and WP for Jalsikka for Actual rainfall 2003 Another example is that of Jalsikka for the year 2002; all other factors remaining the same, the wilting point here is changed as can be seen in the Table As the wilting point is reduced, the available soil moisture has increased. The recharge in all cases is zero and hence the impact could not be qualified as increasing or decreasing. But as seen in the previous graph, as the soil moisture increases, recharge decreases. Table 5.10: Wilting point versus available soil moisture Cla y Sand Sil t FC CP WP FC CP WP Prec PET AET Rech arge Avl soil moisture % % % % % % mm mm mm mm mm mm mm Mm a b c Since the FC, CP and WP influence the available soil moisture and recharge, it will be interesting to see how the soil constants would influence when their values for one layer are the same as when distributed as two layers. In the Table 5.10, row c shows that irrespective of whether the soil is composed of one layer or two layers, the FC, CP and WP are the same in two cases. If there are two layers, then the soil constants are present as average of the two layers as can be seen in row c of Table 5.10.

27 Layers Clay Sand Silt Rainfall* Recharge Avl soil moisture Groundwater Recharge Estimates for Study Villages 251 However, reducing the rooting depth or thickness by half has resulted in increase of recharge by three times and decrease in available soil moisture by half (see Table 5.11). This is because the amount of soil moisture stored in the rooting zone also is reduced by half, as can be seen in the FC, which is also reflected as the increase in the available soil moisture. This response is for the given conditions of water table (which is shallow) and all other climate variables remaining the same. This indicates that rooting depth makes a significant difference in the process of groundwater recharge. S. no. Table 5.11: Impact of Rooting Depth layer on recharge and available soil moisture in Jalsikka C RD FC P WP FC CP WP PET AET mm No. % % % % % % mm mm mm mm mm mm mm mm 2 a b The amount of increase in the recharge is also a function of the depth to water table. The losses where the water tables are shallow include ET losses due to evaporation and transpiration requirements by plants and crops. Deeper water tables would show lower quantum of recharge due to the need for meeting with the moisture requirements of the intermediate layers. Table 5.10, for example, also shows that soil moisture could be available even when the recharge is zero. This however depends upon so many factors which have been discussed in section 1, Chapter 4 as part of literature review; what is important now is to recognise that provided the annual rainfall is above a certain limit, there is a likelihood of recharge. While rainfall lower than 300 mm is most likely to produce no recharge, anything above that level, depending upon the vegetation and climatic factors, may first result in enhancing soil moisture, and more or less after soil moisture saturation, adds to the groundwater table. The threshold limit from rainfall analysis already described earlier is around 640 mm for Ambaredi and around 600 mm for Jalsikka. Rainfall analysis, as described earlier has also indicated that for Ambaredi, an annual rainfall of mm

28 AET, mm Groundwater Recharge Estimates for Study Villages 252 is likely to produce an annual recharge of around 60 mm, which for Jalsikka is between mm of rainfall. Hence, rainfall above 300 mm and below 400 mm for Ambaredi, and above 300 mm and below 560 mm for Jalsikka is likely to add to soil moisture. How do the AET and PET impact on the soil moisture? Analysis of the AET and PET data for the years for both Ambaredi and Jalsikka has shown that AET and PET are both equal for all the years on a year to year basis. The output files of both Ambaredi and Jalsikka from NUT_MONTH also indicate that AET is equal PET when sufficient soil moisture supply is available. Tallaksen and Henny (2004) show that the wet years help carry forward not only soil moisture but also lead to higher groundwater levels. Further, they found that the difference in the AET during consecutive years in 95% of the case is not more than mm/yr. The same has been found for both the clusters of study villages. For example, Figure 5.13 for Jalsikka shows that the variance in AET in consecutive years has been mostly between +50 and -50 mm/yr for all the 102 years from except for a value of 132 mm (difference for the years 1986 and 1987); this is because the highest AET was 1962 mm for the year 1987 as per the CRU data. 150 Year to year change in AET-Jalsikka Year to year change in AET -100 Year -150 Figure 5.13: Year to year change in AET for Jalsikka