Impact assessment of check dam recharge sites in the pappiredipatti Watershed (South India) by using LULC and NDVI Data

Abstract: Agriculture is one of the essential sector in because of the changing phenomenon of weather condition which is further complicated by the interaction of the vegetation with the environment. Scarcity of rainfall causes physiognomic changes can be identied by satellite images through the use of vegetation indices and landuse and land cover changes. The vegetation indices are sensitive to the rate of plant growth as well as to the amount of growth. They are sensitive to change in vegetation affected by moisture stress. Hence, scarcity can be monitored more systematically through remote sensing techniques than the ground based methods of collection of information. The capabilities of Geospatial techniques have been used to demarcate effective sites for monitoring functional characteristics of water conservation structures in the pappiredipatti watershed. In this study, land use/land covers have been used to delineate the existing recharge sites for water conservation measures. Increase of groundwater resource is proposed in the watershed by constructing runoff storage structures like as check dam, percolation tank and gabion structures. The site effective water conservation structures is determined by considering spatially varying parameters land use/land cover and natural difference vegetation index (NDVI) information of the watershed. Geospatial technique has been used to store, analyse, integrate spatial and attribute information pertaining to functional characteristics of land use and land cover within buffering distance in the area of interest of watershed. For detailed changes around recharge area assessment an attempt was made over the watershed. Crop growth area was assessed in and around under forest, barren land and dense vegetation area using temporal normalised difference vegetation index (NDVI) on 1985, 2005 and 2015. They have compared with vegetation area regions, and crop wise area growing were identied. Based on the result obtained from the study area on vegetation cover, buffered around recharge area impact was assessed by NDVI. Decrease in barren land on 2005 by identifying the NDVI threshold over predominantly sparse vegetation growing regions, crop wise area was identied. Further, the water body was estimated from NDVI data of 2005 and 2015 which have been validated for the vegetation assessment. The result of Landsat data use study indicated decrease in dense vegetation area that has leading to wide spread drought condition in the some part of the watershed.


Introduction
Land use and land cover changes can easily indicate growing of agriculture play a signi cant part in the assessment of water harvest in the watershed. Coniferous forests, for instance, ingest more water than deciduous forests, while evergreen forest land and plains area use less water than forests (Bosch and Hewlett, 1982;Brown et al., 2005). Hope et al. (2009) examined the associa tion among yearly river harvests and remotely sensed variations of crop land cover in a great rocky fynbos catchment. Since the consequences of changes, it has chosen that the one of the spectral vegetation index of NDVI with difference a negative and positive connection with the river crop area. Dependent on environment and climatic condition, plant life is a strongly shaping in uence by the surface water process. Robinson et al. (2003) suggested that no important effect of a modi cation in forest cover on mountains and in many basins across N-W Europe, though deforestation directed to an surge in improper ow in more moderate environments of situation (Hornbeck et al., 1993). Land use and land cover change have a straight in uence on the catch ment of water process (Bhaduri et al., 2000;Tang et al., 2005;Ott and Uhlenbrook, 2004;). This impact can be digni ed the change ow of water in river , or forecast by hydrologi cal simulation model is used and combined with land use and land cover change situations of the aspect, it has allows for assess ing the impression of the change on groundwater recharge and discharge of river sites. The in uences of land use change in the context of recharge studies deliberated by various researcher such as Albhaisi et al.,2013,Bhaskar Narjary et al.,2014,Juan Ramón Raposo et al 2013,Tran Van.,2012, Sashikkumar et al.,2017,Kenji jinno et al.,2009,Dams et al.,2008,Hamid Hosseinimarandi et al.,2014Bertrand leterme and dirk mallants.,2011,Stiefel et al.,2009,Vahid Amini Parsa et al.,2016,Ala-aho P et al.,2015, Rana Chatterjee et al.,2009,Marcelo Varni et al.,2013.,Ranu Rani et al.,2009and Oke et al.,2013. An objective of this study to validate the impact of land use and land cover variation on groundwater recharge sites in area of data resulting from Landsat satellite images by using a simulated model (Molusce), for Pappiredipatti watershed, South India. The pappiredipatti watershed was selected because it has face undergone many temporal variation of changes of landuse and soil erosion in recent decades. Check dams are one of barrier on the river site that is making a new environmental cycle of system behind it and non-natural hill slope of living plant. It was imagined that the evapotranspiration would be reduced and it further recharge augmented due to this change in land use.

Study area
Pappiredipatti watershed is situated in the study area are shown in Fig 1. This study area is lies between 78 o 18'0''E and 78 o 30'0''E and 11 o 48'30'' N and 11 o 58''0'northern latitudes in Tamil Nadu, India. The total geographical area of the watershed is 184.55 sq.km of which 67343.75 acres is under cropland. Data on general crop land was collected from eld investigation that may indicated that the sugarcane and coconut tree are the major crop in most of the covering part of area followed by paddy.

Materials And Methods
The main data are used for the mapping of the study area's land use/land cover and NDVI lineaments that are LANDSAT 7 enhanced thematic mapper plus (ETM+) downloaded from U.S. Geological Survey.
The imagery used to comprise of subsets from the original scenes of path 143 and row 52 and path 154 and row 52 of 2005 and 2000 the respectively. Geometric recti cation generated NDVI after applying corresponding sensor calibration coe cients. NDVI images over a period of month is required to get the cloud free NDVI image of the area which can represent the total scenario of vegetation condition in that particular month. The land use and land cover have predicted in the watershed using QGIS software.
LU/LC has been prepared by using unsupervised classi cation and NDVI used in Erdas software. acre. It has obviously illustrated that the simulated LULC map, barren land and agricultural land area are underestimated but the predicted amount of forestland and urban is overestimated. The accuracy of yield is in the assessment process by molusce module using the QGIS software. Resulting from the K values of all well above 0.9 is shown in a satisfactory level of accuracy. The model validation of land use and land cover indicates that 2015 optimum changes occurred in the watershed. Simulation map is a respectable promise with situation of 2015. Therefor it is concluded that molusce models is suitable for prediction of future challenges of planning of LU/LC for sustainable development of agriculture land and urban area.

Validation and Simulations of LU/LC
QGIS is quick and convenient tool for analysis of land cover changes. Molusce is a toolbox to partially automate the process to make a model that can anticipate arrive utilize changes between two decades of period for predict future sustainable development. The raster con guration of land utilize classi cations The Predicted future land utilize changes are utilizing determined model, current condition of land utilize and current elements. Display (Molusce) is a calculation that is utilized for forecast of land utilize changes. State raster is one-band raster where every pixel is alloted with landuse class. Input state raster is one-band raster portraying the past. The Output state raster is one-band raster portraying present prepares the model. Change guide is a whole number one-band raster that stores data about moves.
Classi cation estimations of progress guide are mapped coordinated to move classes.
Simulator module have been performs land use change evaluation process. Introductory state raster contains data about current land utilize classes; gure rasters contain data about illustrative factors.
Model is an indicator that computes change possibilities in the state of the variables and current land utilize. The area impact is accomplished if a model uses neighborhood amid preparing, Simulator considers just broad examples. Test system takes move probabilities from the move framework and gures numbers of pixels that must be changed test system cell demonstrate, go to it starting state raster and component rasters. The model sweeps pixels of the rasters and gures move possibilities of each move class. As result the raster contains the model certainty: the greater is contrast, Simulator develops a raster of the most likely moves: the pixels of the raster are the move class with the greatest capability of move. This raster is utilized amid the following stage for each class. Simulator seeks in the raster of the most plausible moves a required check of pixels with the best certainty and changes the classi cation of the pixels. At least two pixels are close, and afterward arbitrary decision of the pixel is utilized to the one emphasis of reproduction. Approval module is permits to check precision of the recreation of land utilize and land cover.

Effect of land use change
In the watershed, it has been increased irrigation due to check dam construction that strongly indicates the crop yield has impacted positively on socioeconomic. Due to the presence of arti cial recharge structures available water in the wells for the winter crop increased the area under cultivation. The above impact assessment is a strong bene t perceived from the construction of check dams at 500 meter depend upon recharge structures.
The contributions of the different LU/LC classes are for effects on groundwater storing. The consequences of groundwater revive assurance for unmistakably appears from the impact of land utilize change on groundwater stockpiling (Fig.5) The monthly rainfall data were collected from the Public Work Department (PWD) converted into average seasonal rainfall. The data has interpreted for 2000-2015. Taking data, two periods of time has helped to assess the real impact of arti cial recharge structure by comparing low and high rainfall years (Fig 6).
The groundwater level has been due to the recharge from annual rainfall of 971.24 mm and recharge from check dam for growth vegetation. The average rainfall from year 2000 was 990.62 mm higher than the rainfall from year 2015 was 971.24. For comparison reasons in Fig.7 presents the average monthly values of temperature and potential evapotranspiration observed from climatological data. From this comparison, it occurred that the winter temperature was slightly high whereas potential evapotranspiration also increased. Therefore, it was decided reasons due to catchment scale, the mean historical data useful instead of the baseline climatic since the observed differences in and around arti cial recharge structures for further recharge calculation.
In addition, the areal scope and sort of land use classes in the catchment, the appropriation and area of these classes can in uence the groundwater recharge amount use to delineate the year 2005-2015.
Notwithstanding the expansion in fruitless land zone and reduction in woodland zone, the estimations of groundwater energize have contrasted on 2015. Decrease in groundwater revive could be an after effect of the circulation of the bush terrains in the ranges of high energize values; these regions near the seepage arrange in the catchment of the Vaniyar Dam. As the precipitation in the period on 2015 is high whereas changed land use has a signi cant contribution to expand revive. In this way, vegetation contributes to the anticipated increment in groundwater energize. Henceforth, which reproduced the time arrangement from 2005-2015 with the land utilize outline that there is a deliberate extreme of revive increment of 18 % in late year due to the in land utilize little transform from the 2015, which con rms that the signi cant land utilize change unmistakably adds to energize exceptionally.

Proximity of WCS Function by NDVI
NDVI data have generated using the yearly (10 years interval) time composited over the watershed in 1985, 2005 and 2015 by Landsat image. This is presented in Area of interest buffered in the context of Water Conservation Structures (Fig 8) using ArcGIS Software. Buffer techniques can create polygons around input (Vegetation indices) features to a speci ed (500m) distance. Normalized Difference Vegetation Index (NDVI 1985) is in the study area as shown in Figs 9 and 10. With verifying, the temporal variation of NDVI are through years 1985 NDVI over the region that indicates low crop condition whereas 2015 indicates high crop condition. Based on the prepared GIS layers of different land use, the total crop area has been extracted from the satellite. The major crops were classi ed by temporal variation. From the data, reduction in crop area is observed in 1985 when compared with 2015 because of the prolonged dry spell that occurred in the month of july-august 2015 which is critical for crop sowings. Barren land (0.33-0.10 in whole area, around check dam -0.07 to 0.10 and percolation pond -0.03 to 0.10), Sparse vegetation 0.10-0.50 compared to recharge sites is 0.10 to 0.43. Area of interest buffered (NDVI 1985) in the context of water conservation structures is dominant of sparse vegetation than dense vegetation. Changes in the vegetation condition are weighted for average with NDVI and water de cit values obtained from water planning procedure for crops comparison. The results indicated that the paddy dominated mandals de cit water in July the NDVI value has fallen 0.37 to 0.17 and due to continuous de cit values the NDVI continuous to fall down till September. It is observed that due to constant water de cit over kharif the NDVI has fallen in August and September even though NDVI is good july. In mandal due to high de cit in the month july the NDVI is fallen down in August and September. Scatter plot (Fig 15) showing In the present study, the relative abundance of the individual Land use/land cover categories to represent their respective patch frequency while the richness represents the number of LU/LC categories considered in the present study which is equal to eight. The Shannon's index of diversity would be primarily determined by the relative abundance of the eight LU/LC categories such as in 2005(H) are ranges 1.45 to 1.29 and after images 2015(H) are ranges 1.68 to 1.5,evenness in prior images ranges between 0.75 to 0.62 after ranges 0.80 to 0.55 in the study area. The highest abundance of the individual is indicating their functional perform in the table 1  Simpson's Index of Diversity are ranging between 0 and 1, but now, the greater the value, the greater the sample diversity. In this case, the index represents that randomly selected from a two different land use and land cover. Simpson's Reciprocal Index starts with 1 as the lowest possible. This table 2 would represent one individual dominant land use and land cover. The higher the values indicates the greater the diversity.

Field photographs of inspected area
Siltation in this area is characterized by dry and heavy monsoon rainfall and high erosion because the high sediment load to occur at the check dam site originates from the hillside of ows near check dam carrying high ne-sediment load as shown in Fig.17. However, the sediments carriers of sediment which carried are coarse and ne accumulated at the check dam site. Though the upstream gradient is low to moderate, the velocity of water during the monsoon is high and this is evident by the large amounts of sediments and boulders in the behind of check dam.

Conclusion
This study clearly indicates to be useful for planning and management for sustainable watershed development through natural and arti cial conservation structures by response of land use (LU) and land cover (LC) to increase vegetation. It is necessary for the land use planner retrieval data about past, present and future to make appropriate decision regarding land use for future challenges. Landscape alternation in the urban area is more signi cant as compared to past in this watershed. Deforestation and human impedance are more essential dangers to the exceedingly rate biodiversity asset around there.
The simulated land use map is bene cial for water conservation where exact site selection for well development. In the watershed has been carried out to estimate water de cit and found good relation with reduction in NDVI. It is also observed that there is reduction in total crop area as well as in the dynamism in the vegetation in 2005 when compared with 2015 and it indicates de ciency of rainfall in watershed. The detailed analyses of NDVI clearly are indicated drought condition assessment in the watershed for development of agriculture pattern and predict future vegetal cover.

Declarations Competing Interests
The authors declare no competing interests