Impact assessment of indigenous field-level water storage structure (diggi) on agriculture in the great Indian desert

In the arid region of Rajasthan, India, it is very often a challenge to store rain/surface water for year-round use by human and livestock. The inhabitants of this desert area have developed several water storage structures, which they used to construct based on their indigenous knowledge of local terrain conditions. Recently, farmers living in the lower command areas of Indira Gandhi canal have constructed micro-farm water storage structures, called diggi, in their cropland. They store allocated canal water in diggi to timely utilize it as per irrigation requirements, and thus, done away with problems of deficit and untimely canal water supply. This impact assessment study, carried out in Poogal tehsil of Bikaner district, analyzed high-resolution satellite images for 2 years and used a geographic information system to quantify diggi structures and studied the diggi-growth interactions with changing rainfall pattern as well as land use/land cover. Results showed that during 2018, about 3243 number of structures were constructed compared to 241 existed during 2004–2005. The mean annual rainfall increased by 30% and crop area by 12,772 ha during the same period that included 7.86% increase in irrigated and 2.98% in rainfed croplands. Thus, diggi-based water management has been immensely helpful in the perspective of irrigated agriculture in the desert region.


Introduction
As we move into the next millennium, the biggest and fundamental concern for the society will be on how the people manage the available water. As per estimates made by various world water forums, by the year 2030, one-third of world's population will feel the heat of water scarcity and majority of such population (about 2.1 billion) will be from world's dryland region (UNESCO, UN-Water 2020). India has about 31.7 million ha (m ha) land under hot arid zone and a major proportion of such land (62% or 19.6 m ha) lies in the Great Indian Desert region situated in Rajasthan. The arid region of Rajasthan, which is bounded by aridity index values between −59.5 in the east (in Sikar district) and −88.9 in the west (in Jaisalmer district) comprises 12 districts. There is a wide spatial variability in the rainfall distribution that influences the availability of water resources. This is primarily due to low (100-600 mm yr −1 ) and erratic rainfall (coefficient of variation ranging from 20 to 65%) coupled with high potential evapotranspiration (~2000 mm yr −1 ). The annual rainfall is the lowest (100-150 mm) in the west and the highest (500-600 mm) in the east that helps about 75% of desert dwellers to continue agriculture at least once a year. With scarce and unreliable rainfall situation, status of groundwater and surface water is non-uniform and highly variable in space and time. With a long-term average annual rainfall of 317 mm, the total annual estimated groundwater availability in western Rajasthan would be 3351 million m 3 as against the total annual water demand (draft) of 7790 million m 3 in the year 2020 with 221% stage of groundwater development (Goyal et al. 2009). However, people extract the groundwater indiscriminately through dug wells and tube wells and use it for drinking and agricultural purposes. Due to over-extraction, groundwater levels have been fast-depleting in 9 of total of 12 arid districts of Rajasthan, and Bikaner is one of those districts. Surface water in Bikaner district is handicapped as there is hardly any major river and as such 1 3 284 Page 2 of 14 freshwater received during rains is stored in various forms of natural and man-made structures. Thus, common people of this region struggle hard to collect water, conserve it, and use it efficiently for various purposes. Such activities are very important for desert areas, as (i) this region is too dry for conventional rainfed agriculture, (ii) it is home to about 27 million people, and (iii) it is a region where demand for water is going to increase because of sprawling demands from agriculture, industry, and service (tourism) sectors. Therefore, people of this region always think about alternate ways and means of managing the available surface water.
Management of surface water is one of the daunting tasks for people living in western Rajasthan as per the requirement of growing human (27.09 million, 2011 census) and livestock populations (27.6 million, 2019, 20th livestock census, Source: http:// anima lhusb andry. rajas than. gov. in/ lives tock_ census. aspx). Therefore, local people of this region have constructed a number of traditional as well as improved surface water harvesting structures like tanka, kund, nadi (village ponds), khadin (a runoff harvesting system) and baoli (Narain and Khan 2000). A tank, locally called as tanka, is a man-made structure for individual household and also serves the community having higher storage capacity ranging from 5000 to 600,000 L of water. A tank of 21,000 L capacity is found sufficient to meet the drinking water requirement of a family with 6 persons for round the year (Goyal and Vittal 2008). A baoli is regarded as a community well as such wells were built by banjara or mobile trading people for drinking water in Rajasthan (Hussain et al. 2014). The nadi or village pond is a common rainwater harvesting structure representing one of the oldest practices in desert areas. Such structures are utilized by about 42% of population in arid zone of Rajasthan. These ponds store runoff water from adjoining natural catchments during the rainy season. Most of nadis have a capacity of 1200 to 15,000 m 3 and store water for a period ranging from 2 to 8 months in a year. Khadin is a unique system of water harvesting, moisture conservation, and utilization. Alone in Jaisalmer district, there are more than 500 khadins (Goyal et al. 2018). The main feature of khadin is a very long (100-300 m) earthen embankment built across the lower hill slopes lying below gravelly uplands. This system has the dual benefits of harvesting rainwater on farmland and subsequent use of this water-saturated land for crop production. The water-holding capacity of khadin soil ranges from 200 to 250 mm m −1 depth. Reservoirs are another common form of water storage structures in higher rainfall zones mainly in the east adjoining to Aravalli hills, which receive more than 350 mm average annual rainfall. The Jaswant Sagar dam built on the Luni River near Bilara, Sardar Samand dam on the Guhiya River to the north of Pali, the Hemawas dam on the Bandi River at Pali, the Bankli Talav on the Sukri River near Bhadrajun, and the Jawai dam built across Jawai River are historically important and are still relevant for region's agriculture and drinking water. Besides, this part of the state also gets the benefit of two big canal projects, IGNP (Indira Gandhi Nahar Pariyojana) and NCP (Narmada Canal Project). The IGNP is a large surface water transfer project constructed during the 1960s in the heart of desert area with an objective of using this water to transform the extensive sandy wasteland of western part of Rajasthan into the greenery. Within Rajasthan, the canal is 445 km long passing through the sand dunecovered terrain in 7 of 12 arid districts. The IGNP Stage-I is restricted to Sriganganagar, Churu, Bikaner, and Hanumangarh districts and the Stage-II to Jaisalmer and Barmer districts. The entire project area serves a vast Culturable Command Area (CCA) of 19.63 lakh ha land of the region (Sharma et al. 2010;Shrivastava et al. 2013). The NCP caters to the needs of people of Barmer and Jalor districts.
Recently, Diggi has been added as another water storage structure by farming community in the IGNP command area in western Rajasthan. In India, farm pond structures similar to diggi in Rajasthan are very common in the eastern region where normal annual rainfall is much higher than Rajasthan. The 50-year (1951-2000) normal annual rainfall varies from 500 to 1000 mm in Andhra Pradesh and Telengana, and 1001-2000 mm in West Bengal, Odisa, Bihar, and Jharkhand states (https:// mausam. imd. gov. in/ imd_ latest/ conte nts/ rainf allin forma tions. php#). There have been reports ) of positive impacts of such structures on fisheries and livestock rearing, crop diversification, and improvement in water availability. The basic objective of farm pond structure is one way to deal with water scarcity. In Rajasthan, even under an arid climate, low and erratic rainfall, and poor water resources conditions, farmers have started constructing such structures in a large number in the canal command area. Over a period of time, there has been a situation of water scarcity and uncertainty in the timely supply of IGNP canal water that creates problems in the tail end of CCA. Under such conditions, farmers of the region have evolved a new way to conserve and manage canal water in their farm ponds called diggi. A field photograph of diggi structure is shown in Fig. 1.
During a semi-detailed survey of this area (2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018), it was found that such structures have been increased to a large extent. This study has been undertaken with three objectives: (i) to quantify the spatial distribution of diggi structures, (ii) to explore changes in rainfall patterns, and (iii) to understand whether such indigenous field-water management systems have any impact on the growing irrigated agriculture of the region.

Study area and data description
This study was carried out in Poogal tehsil of Bikaner district in Rajasthan, India (Fig. 2). As per the 2011 census, Poogal tehsil has a geographical area of 2616.23 km 2 . The tehsil is located within the desert area of the district where 58% of the area is covered by sand dunes and 33% area by sandy plains (Moharana et al. 2013). The western, southwest, northern, and northeast parts of the district are largely covered with 12 to 20 m high sand dunes. The district receives a scanty annual rainfall that varies from 300 mm (in the east) to 180 mm (in the west). The surface water resources of the district are rather meager, due to erratic and low rainfall, large infiltration rate of the upper sediments and practically having no river system. Under such situation, the segment of IGNP (Indira Gandhi Nahar Pariyojana) passing through the district is the lifeline for people, livestock, and agriculture. As per district agricultural statistics of 2016-2017, out of total 1.59 m ha cultivated area in Bikaner, 0.46 million ha area was net total irrigated. The IGNP canal contributes to 0.20 m ha or 44.51% of total net irrigated area and rest (55.49% area) is irrigated through dug wells and tube wells. Between 2006Between -2007Between and 2016Between -2017, total sown area during kharif season (July to October) has increased from 1.31 to 2.07 m ha. During rabi period (October to March), total sown area has increased from 0.28 m ha (2006)(2007) to 0.50 m ha in 2015-2016 (http:// www. agric ulture. rajas than. gov. in/). Poogal tehsil is dominantly a rural town, and the major occupation of the local people is overwhelmingly agriculture (in ~90% area). As per the 2011 census, the study area is inhabited by 67,163 people, 11,977 numbers of households with a population density of 26 inhabitants per km 2 against 78 per km 2 of Bikaner District. Paralleling the rise in human population, there was a corresponding rise in livestock population by ,7.1 % as well as an expansion of croplands (i.e., from 47.17% in 2005-2006 to 67.43% in 2015-2016), while the extent of grazing lands dropped from 1.70 to 1.62% during the same period.
In this study, annual rainfall data of the study area (Bikaner District) are collected for a period of 20 years (2000-2019) from website of Water Resources Department of Rajasthan (https:// water. rajas than. gov. in). This study utilized satellite imageries from Indian Remote Sensing (IRS) Resourcesat-1, Linear Imaging Self-Scanning (LISS-IV) satellite images of 5.8 m spatial resolution (two imageries), IRS-LISS-III images of 30 m spatial resolution (one imagery), Sentinel 2A optical image (10 m spatial resolution) of ESA (European Space Agency), and Google Earth Images. The IRS images were purchased under projects, while the Sentinel 2A imageries were downloaded from the website (https:// earth data. nasa. gov/ earth-obser vation). Moreover, data related to crop statistics such as cropped areas, yields, and productivity of the major crops in the study area were downloaded for 2007-2008 and 2017-2018

Analyzing rainfall variability
Basic statistical properties such as mean, standard deviation, coefficient of variation, skewness, and kurtosis of annual rainfall data are computed. Bar charts of annual rainfall data for the period 2000-2019 are plotted to visualize the rainfall variability and a linear trend is fitted. A box-whisker plot of the annual rainfall is also plotted to depict five important properties of rainfall time series, i.e., minimum, maximum, 25th and 75th percentiles, and outliers/extremes (Machiwal and Jha 2012).

Identification and mapping of diggi structures in geographic information system
In this study, diggi structures are identified and delineated in geographic information system (GIS) using the satellite images of the years 2004 (IRS-LISS 4) and 2018 (Sentinel-2A). These micro field-water storage structures are clearly visible in the remote-sensing satellite (optical) imagery having very high spatial resolutions of 5.8 and 10 m, for IRS-LISS 4 and Sentinel-2A, respectively, and are verified from Google Earth Images (Figs. 3 and 4). These satellite images ease the process of identification and mapping of diggis due to their high spatial resolution of 10 m, and thus, diggi features comfortably visible on the imageries are precisely mapped on the GIS platform. Three features of diggis that helped in their quick identification are (i) size of diggi structures, (ii) their shape (circular or rectangular), and (iii) their topological association with canals and croplands. Once identified, diggi structures are mapped through on-screen digitization from false-color composites (FCCs) of satellite images for the years 2004 and 2018 using ArcGIS (10.5) software. Locations of the delineated diggi structures are verified from the Google earth images.

Understanding water storage-cum-application system of diggis
To understand the water storage-cum-application system of diggis, information and details were collected from the beneficiaries through a survey regarding diggi operation (material used for construction, side slopes, storage capacity, lining material, if any, time taken in filling and emptying diggi completely, evaporation loss, type of pump with horsepower and discharge, etc.), canal runs (daily operation hours, monthly period, etc.), and crops grown (name, irrigation method, number of irrigations, etc.).

Mapping of land use/land cover
Land-use/land-cover (LULC) maps of the study area are developed for 2 years, 2004 and 2018, using satellite imageries of IRS Resourcesat-1, LISS-III, and Sentinal-2A, respectively. The objectives of preparing these maps are to understand the spatial distribution pattern of irrigated croplands around the farm pond structures and to observe the distribution of other types of LULC in the area. For LULC mapping, we used the national land-use/land-cover classification system developed by the National Remote Sensing Centre, Hyderabad. As per this classification system, there are 5 Level I classes, 11 Level II classes, and 15 Level III classes. It is worth mentioning that LISS-III were used for mapping land-use/land-cover classes, while for mapping the diggi structures, IRS-LISS-IV and Sentinel-2A have been used. Supervised classification system followed by manual modification techniques was used for classifying the landuse/land-cover types.

Cause-effect analysis of diggi structures with changing rainfall and agricultural statistics
In the arid climate of the study area, water remains available in scanty quantities for agriculture and allied sectors. Perhaps, this is the reason that rainwater harvesting is generally not considered as a dependable source of water supplies in the arid regions. Hence, the motivation behind the construction of diggi structures in the area is explored by examining gradual trend in rainfall series by applying the Mann- Kendall test (Mann 1945;Kendall 1975), which is a non-parametric statistical test that is widely used for testing trends in hydrological studies. In this study, trends and patterns of rainfall are analyzed to identify whether the region received any major rainfall events during the study period. If rainfall quantities are found increasing, then it would help in investigating whether the increase in crop area, if any, also reflects the contribution of rainfall. Details of Mann-Kendall test are available in the literature and may be found in Machiwal and Jha (2012). In addition, Sand Dunes, Irrigated croplands t test is employed to evaluate abrupt changes in the mean of the annual rainfall over the years.

Exploring changes in mean rainfall by applying t test
The parametric t test is used to examine whether the change in the mean annual rainfall between the first (2000-2009) and second (2010-2019) halves of the total 20-year series is statistically significant or not. It considers that the annual rainfall series is uncorrelated and normally distributed with mean μ and standard deviation σ, and the two subseries have equal variances. To apply the test, the whole rainfall series x t (t = 1, 2, …, n) is first divided into two subseries: (i) first subseries x i (i = 1, 2, …, n 1 ) of size n 1 has a mean 1 , and standard deviation σ, and (ii) second subseries x j (j = n 1 + 1, n 1 + 2, …, n) of size n 2 has a mean 2 and standard deviation σ such that n = n 1 + n 2 . Using the variance rule of thumb, the ratio of the larger variance to the smaller variance (1.30) was less than 4, and thus, an equal-variance or pooled t test was used in this study. The test statistic is defined as (Snedecor and Cochran 1980) where x 1 , x 2 , s 2 1 , and s 2 2 are the estimated values of the means and variances of the first and the second subseries, respectively. If the computed value of the test statistic is greater than its critical value (at n-2 degrees of freedom and 5% level of significance), then the null hypothesis that the mean of two subseries belongs to the same population is rejected.

Evaluating impact of diggis on changes in land use/land cover
Sand Dunes, Irrigated croplands

Temporal characteristics of rainfall
The 20-year mean annual rainfall in the study area is found to be 240 mm with value of standard deviation as 104 mm, which indicates that the area is rainfall deficit or drought prone. Value of the coefficient of variation for the annual rainfall is 43%, which suggests that the annual rainfall is highly variable in the area. Positive value of skewness (0.52) indicates that rainfall distribution is skewed to the right and close to zero. Likewise, positive value of kurtosis (1.13) indicates that the data distribution of annual rainfall is heavy-tailed with a sharper peak in comparison to the normal distribution curve. Similar findings are reported for semi-arid regions of Rajasthan and Gujarat (Machiwal and Jha 2017;Machiwal et al. 2021).
Rainfall is considered as the most important deciding factor for agricultural planning in desert region. Bar charts of the annual rainfall in the study area showed the occurrence of a few good rainfall totals ranging from 300 to 500 mm during 3 years, i.e., 2005, 2008, and 2015, and also very low annual totals of 34 mm during a severe drought in 2002 and about 150 mm during 2004 and 2009 (Fig. 5a). The fitted linear regression line depicts an increasing trend with a magnitude of 4.62 mm year −1 in the annual rainfall. However, the linear regression trend is not perfectly fitted as the value of the coefficient of determination (R 2 = 0.07) is close to zero (Fig. 5a). Box-whisker plot of the annual rainfall is shown in Fig. 5b, which reveals that the median value of the annual rainfall is 232 mm in the area. It is further seen that the median value of the rainfall is not exactly at the center of the box, and lengths of the upper and lower whiskers are also not perfectly equal to each other. This suggests that the rainfall distribution is slightly non-uniform, which is further confirmed by the presence of one outlier on the upper side that makes the distribution skewed to the right. This finding is in agreement with that obtained from the skewness value. Hence, annual rainfall has large variations but shows no significant increase in the area over the past 20 years.

Spatial mapping and changes in distribution of diggis
The diggi structures digitized on-screen using satellite images of two time periods, 2004 and 2018, are shown in Figs. 6 and 7. Mapping of diggis revealed that compared to the year 2004, there appears a greater number of diggis during 2018 and higher density of diggis was found in the central, north, and northwest parts of tehsil mostly in Kheersar, Poogal, Karanpura, and Ramai villages. A comparison of statistics for two diggi features, i.e., number of diggis and area occupied by diggis, for the two time periods reveals that diggi structures increased by 13 times from 241 during the year 2004 to 3162 numbers during the year 2018 (Table 1) Rainfall pattern in the study area over 2 decades (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019) depicted at least five droughts including the worst drought of 2002, which was the severest not only in Rajasthan but also in the history of India having its impact on 56% of the geographical area and the livelihoods of 300 million people in 18 states (Samra 2004). Under such conditions, farmers of this region could harvest a single annual crop of pearl millet. However, much of the problems of water requirement have been solved due to the IGNP canal system. This is the only available source of assured water supplies to meet the water demands of people living in the extreme arid districts of Bikaner, Jaisalmer, and Barmer. With time, farmers of this region have availed the benefit of canal water and as an impact, some important transformations have been taken place in the agricultural sector such as (i) many of the areas having low to medium sand dunes have been leveled for cropping, (ii) mono-crop areas have changed to double cropping, (iii) farmers are now cultivating highly water-intensive crops (from traditional cereal crops to more profitable horticulture crops), and (iv) there is a shift from traditional rainfed systems to irrigated agriculture. All these activities require additional water, thus prompting people to think of alternate ways like diggi system to store their quota of allocated water of Indira Gandhi canal. Water is transferred to diggis to be used later on as per requirement (Goyal et al. 2019). Farmers pump out water from diggi using electric motor and then apply to field by micro-irrigation devices such as sprinklers and drip systems. The Government too has recognized the importance of diggi in these areas and came out with a standard design of diggi (110 feet × 110 feet × 10 feet size with 18 feet side slope) to store about 25 lakh liters of water. Farmers have also prepared a system connecting the diggis with photo-voltaic solar panels to run the electric pump for operating sprinkler and drip systems (Tewari 2012). Government of Rajasthan is also providing 50% subsidy for constructing the diggis (Amarasinghe et al. 2008). To reduce the construction cost, these diggis are lined with high-density polyethylene (HDPE) sheets of 300 to 500 micron thickness. It is understood from the above discussion that the requirement of additional water to meet irrigation

Water storage and supply system of diggis
Diggis in the study area are constructed to store both canal and groundwater irrigated areas with the purpose of utilizing stored water as and when required in adequate amounts and at right time. This study focused on the situation of diggis in the canal command area only. Diggis are constructed on the ground with loose soil excavated from nearby area.
The soil is compacted and a side slope of 1:1 or 1.5:1 is maintained for the stability. The diggis are lined with highdensity polyethylene sheets of 300 to 500 micron to make them cost-effective. The storage capacity of a diggi depends on individual farmer's landholding. However, the state government recommends the minimum capacity of 4 lakh litres with 0.5 ha cultivable land. For the higher land holdings, the diggi is designed to store water required for providing at least the single irrigation of 5 cm depth in the field. The state government provides 75% subsidy of unit cost with the maximum amount of 3 lakhs. In canal commands of the study area, water is delivered through warabandi method, which is a rotational system of water supply for making an equitable distribution of the available water in an irrigation system. In this system, the canal water is distributed in turns fixed according to a pre-determined schedule specifying the day, time, and duration of supply to each irrigator in proportion to the size of landholding in the outlet command (Malhotra 1982). In the study area, canal water flows for a week and the canal remains dry for the next week, and thus,  the farmers get their quota of canal water for irrigation at fortnightly intervals. The canal runs for 4-5 h in each turn and diggis in-turn get filled in 4-5 h. Then, the water stored in the diggis is pumped using solar-powered submersible electric pumps of 3 horsepower and discharged at 4-6 L per second to croplands using micro-irrigation methods such as sprinklers usually having 20-25 nozzles. The pumps in the diggis are kept in a floating position (minimum 6-inch from the bottom) to reduce the suction head. The diggi is completely emptied in about 8 h. The water in a diggi can remain for the maximum period of 15 days, i.e., before the next turn of the canal delivery. Under the arid climate of Rajasthan, the amount of water lost in the winter season due to evaporation can be as large as 50 mm at the mean daily rate of 3.33 mm. In the study area, wheat, mustard, and gram crops are irrigated with canal water using the sprinkler method of irrigation.

Land-use/land-cover maps
Nine classes of land-use/land-cover (LULC) units are identified and mapped in the study area. The LULC maps for the years 2004 and 2018 are shown in Figs. 8 and 9. In the year 2004, it is seen that the sandy/dune LULC class dominates in the area encompassing 184,924 ha (70.68%) area followed by fallow lands (33,851.60 ha, 12.94%) and area under irrigated agriculture (25,660.10 ha, 9.81%) ( Table 2). These three LULC classes cover 93.43% of total study area in the year 2004. Similarly, in the year 2018, sandy/dune LULC class covers 169,938 ha (64.96%) of the study area followed by irrigated agriculture (46,221.30 ha, 17.67%) and fallow lands (16,012 ha, 6.12%), and hence, these three classes have a spread over 88.75% lands. Agricultural lands including irrigated croplands, un-irrigated croplands, and fallow lands constituted 25% of the study area in the year 2004, which got increased up to 29.02% in the year 2018 (Table 2). These results indicated a significant increase in total crop area and the majority of such lands were irrigated types.

Interactions among rainfall-diggis-agricultural statistics
The calculated test-statistic (z) value of the Mann-Kendall test is obtained as +0.91 (p-value = 0.36), which indicates  a positive trend in annual rainfall series of the study area. However, as the calculated z-value is less than its critical value (+1.96), the increasing trend is not statistically significant at 5% level of significance (p-value > 0.05). It is further observed from Sen's slope estimation test that the magnitude of the non-significant increasing trend is 3 mm year −1 . The value of ts-statistic of t test is computed to be 1.38, which is less than its critical value (1.73) at 5% level of significance but more than the critical value (1.33) at 10% level of significance. Hence, it is statistically evident that the rainfall means during two 10-year periods, i.e., 2000-2009 and 2010-2019, are significantly different (p-value < 0.10). Also, as the mean annual rainfall for the period 2010-2019 (271 mm) is higher than the mean annual rainfall for the 2000-2009 period (208 mm), it is apparent that the amount of annual rainfall has been increasing over the recent years in the area. This shift in the mean annual rainfall is depicted in Fig. 5. Similar findings of increasing rainfall quantities in arid regions of Rajasthan and Gujarat are reported in the literature (e.g., Kharol et al. 2013;Machiwal et al. 2016;Meena et al. 2019;Machiwal et al. 2019). The rising amount of annual rainy events in the Indian arid regions over the last 1 or 2 decades augmented the surface water availability in canals. The rising rainfall quantities have played a significant role in large increase in number of diggis in the entire western arid region of the country that stores surplus quantities of canal water at times of their availability, and the same is used later to timely irrigate the crops. A few farmers in the area have been adopting the high-yielding crop varieties to enhance the crop production as well as productivity.
Since the sole objective of diggi-based water management is to provide timely irrigation to crops, the change in LULC of the study area is also detected over the same period, i.e., from 2004 to 2018. The LULC changes have been analyzed by preparing a matrix. The table shows the change detection matrix list land-cover classes of 2018 image in the columns and land-cover classes of 2004 image in the rows (Table 3). Based on this matrix, it is shown that mainly large sand dune areas have been converted to plantations/ shelterbelts (20,843.8 ha) and agricultural use (8681.83 ha) from 2004 to 2018. Parts of scrublands (2120.95 ha) have been converted to agriculture under irrigated types. Fallow lands have decreased and are now converted to agriculture (10,432.65 ha). Parts of irrigated croplands are now utilized under plantations and shelterbelts.
It is learned that the rainfall in the study area has been fairly increasing; though there have been drought years. Thus, the increase in crop area mainly under irrigated lands is due to expansion in croplands facilitated by timely utilization of stored canal water through diggis and slightly increasing rainfall that have enhanced the canal water potential in the area. The area under sand dunes in the area has also decreased by 5.72%. Because of the storage water in diggis, people have also stopped the earlier practice of land fallowing. Thus, fallow land areas have decreased by 6.82%. Similar findings have been reported in earlier studies (Kar 2014). It was also found that most of the new irrigated croplands were diggi centric. Major Kharif crops of this region are pearl millet, moth bean, cluster bean, and sesame under rainfed conditions and groundnut, cotton, and wheat under irrigated conditions. During the winter season, chickpea, mustard, and wheat are grown as the major crops. Comparative data for a decadal change in crop area, crop production, and yield for 2007-2008 and 2017-2018 in Bikaner district are shown in Table 4. The results showed a significant increase in crop area and irrigated area under each crop  (Table 5). The total production and higher yield of crops are also achieved, which are due to important contributions of management factors like timely application of water at different crop growth stages.

Conclusions
Diggi is an indigenous micro-hydrological structure that has become a quite popular strategy for water management at the field scale in the canal command area of desert districts in Rajasthan, India. Satellite image-based interpretations of study area indicated significant increase of 13 times in the number of diggi structures between 2004-2005 and 2018, thus, indicating people's requirement and magnificent adoption of such structures by the farming community within the command area of IGNP. It is apparently seen that rainfall has been fairly rising in the arid lands of the Great Indian Desert, which have positive implications for improving potential availability of canal water in this region. As an impact, local farmers are now able to follow crop calendar, to decide irrigation schedule, and decide choice of cropping pattern. The perpetual problems of erratic and limited electricity supply and disruption in irrigation activities have also been minimized. Farmers are able to expand their irrigated crop area. The change detection matrix for two periods indicated that large areas under sand dunes, scrublands, and fallow lands have been converted to agriculture.The water use efficiency has also improved by bringing additional area under cropping due to water saved by sprinkler or drip systems in comparison to flood irrigation. However, study also found a few limitations of this system; the water in diggis is also prone to evaporation losses under high desertic conditions. Farmers will have to judiciously and timely utilize the stored water;  http:// agric ulture. rajas than. gov. in/ conte nt/ agric ulture/ en/ Agric ulture-Depar tment-dep/ agric ulture-stati stics. html (2007)(2008)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018) S otherwise, failure and negligence will incur the huge loss of precious water. It is also suggested that effort should be made to increase vegetation around diggi that will improve the micro-climate, reduce ambient temperature, cause less evaporation losses, and also reduce siltation of canals.