2.1. Study Area
The current study was carried out in lower Gangetic flood plain of West Bengal using geospatial technology, where wetland ecosystem is vulnerable. We selected Malda district is also a part of Diara region, where the wetlands play a vital role for maintaining landscape stability. Geomorphologically, the study area belongs to the active floodplain as it occupies more than half of the region’s area occupies marsh, inactive floodplain, and levee and river islands. The Malda district lies between 24° 40ʹ 20ʺ N to 25° 30ʹ08ʺ N and 87° 45ʹ 50ʺ E to 88° 28ʹ10ʺ E (Figure 1). It includes 15 blocks and 2 sub-divisions, such as Chanchal and Malda Sadar. The total geographical area of Malda district is nearly 3652.75 km2. The total population of the district is 3699312 with a population density of 1013 persons per sq kilometer in 2011. The major rivers in the district are Ganga, Punarbhaba, Mahananda, Pagla, Kalindri, Tangon, Fulahar etc. The river Ganga has enough evidence of lateral migration (Mukherjee and Pal, 2017). With the shifting of river course, it creates numerous ox-bow lake, cut-off channels, beels, etc. which are no longer direct link with parent river (Singh et al., 2019). Such water bodies have evolved into riparian wetlands that are vital to the floodplain ecosystem as well as the human wellbeing. Such dynamics physical and socio-economic aspects were the main attraction of researchers and that’s why we selected Malda district as our study area.
2.2. Data and Methods
2.2.1 Data resources and processing procedures
Systematic procedures were done to demonstrate the health conditions of wetland ecosystem in Malda district. The systematic procedures involve- preparation of block level vector map of Malda district and tabulation of the health indicator data. The block level administrative map of Malda district was prepared by using the administrative atlas of West Bengal; the statistical data incorporates population density, percentage of cultivation area, urbanization rate is gathered from the Census of India 2001 and 2011. The road network map was prepared from the Survey of India topographical sheet and same has to be updated with the satellite data. Finally, the digital database of road network is used for the calculation of road density. The patch richness (PR), patch density (PD), largest patch index (LPI), landscape diversity index (LDI), wetland degradation rate was calculated by using a total number of six Landsat satellite images in the year 2001, 2011 and 2018. The Landsat pace borne satellite images were downloaded via the web site https://earthexplorer.usgs.gov/. The description of satellite data has been given in Table 1. The ecosystem function value index was used for the economic valuation of ecosystem services provided by the wetlands (Zhang et al., 2015).
2.2.2 Wetlands ecosystems health assessment system
The PSR model was common and popular method along the numerous research studies (Mi et al., 2005; Sun et al., 2016, 2017). The PSR model supported the cause-effects relationship, which can assess the human pressures and exercises cause to deterioration to ecological environment. The PSR model differentiates the indicators into three shorts of categories- (1) The ‘pressure’ factors are the negative factor influences to the deterioration of wetland health; (2) The ‘state’ factors demonstrate the structural integrity and functions status of the wetlands; (3) The ‘response’ indicators are quantifies its vigour, versatility and environmental services. Based on the availability and flexibility of data source, we selected ten indicators for the development of wetland ecosystem health index (WEHI) in Malda district. The analytical hierarchy processes (AHP) techniques have been employed for assessing the weights of the indicator. Detailed methodology is given below with a flow chart (Fig. 2).
2.2.3. Wetland extraction and classification
We have used six Landsat satellite data for the extraction of wetlands on 2001, 2011 and 2018. After the essential image corrections guided by the standard procedures, we calculated Normalized Difference vegetation index (NDVI), Modified Normalized Difference Water index (MNDWI), Normalized Difference Pond index (NDPI) using the Arc GIs 10.3.1 software.
Normalized Difference vegetation index (NDVI) has been calculated based on the Eq. 1
NDVI = (NIR – RED) / (NIR + RED) ……………..EQ.1
NDVI is commonly used for determining different aspects of plant characteristics. The NDVI is first defined by Rouse in 1973 (Rouse et al., 1974). In this study, we have found that the NDVI value ranges between – 0.674 to 0.658. Basically, the positive value of NDVI represents green vegetation and the negative value formed from clouds, water and snow.
Modified Normalized Difference Water index (MNDWI) has been calculated based on the Eq.2
MNDWI = (GREEN – NIR) / (GREEN + NIR) (2)
The MNDWI has been used where the water bodies are mixed with built-up area (Xu, 2006). In this study, we have found that the MNDWI value ranges between – 0.566 to 0.893. The positive values of MNDWI denote water features and the negative values indicated the non-water feature.
Normalized Difference Pond index (NDPI) has been calculated based on the Eq.3
NDPI = (MIR – GREEN) / (MIR + GREEN) (3)
Lacaux (2006) first described Normalized Difference Pond index (NDPI) for the identification of ponds. In this study, we have found that the NDPI value ranges between – 0.986 to 0.566.
After the calculation of above indices, MNDWI is greater than zero from this maximum water features. Next, we have combined NDVI value of greater than zero and NDPI value of less than zero. This condition can separate water features from vegetation. The hybrid image provides better result.
Based on the extracted surface waterbodies, the wetlands are classified as natural lakes, natural ponds, ox-bow type or cut-off meander, natural waterlogged, natural riverine and man-made ponds. The classification scheme was considered based on the Indian Space Research Organization (ISRO) National Wetland Atlas report in the year 2011.
2.2.4 Indicator system establishment
The systematic studies involve selection of the indicator and standardize them for the computation of a score which is described the actual continuity of the study area on a specific topic. Development of wetland ecosystem health index of Malda district, we select ten health indicators, where four pressure, four state and two response indicators. Table 2 shows the detailed list of indicators. Max – min normalization method has been followed to unify the indicators. The scale of the data ranges between 0 to 1.
The ecosystem health pressure factors describe the intensity of human activities due to extensive conversion of wetlands to usable land within a time span. Based on literature study, we have selected population density, percentage of cultivated land, urbanization rate and road density for assessing wetland ecosystem health in Malda district. The population density is calculated as the persons per square kilometer of area in the year 2001, 2011 and 2018. The road density is calculated per square kilometer. The urbanization rate is calculated by subtracting percentage of urban population from succeeding year. The percentage of cultivated area is calculated as the ratio between cultivated area and total geographical area of each block.
The state factors are considered as wetland under the human pressure. Based on the previous studies (Sun et al., 2017; Jia et al., 2015; Sun et al., 2016), various indices have been employed to define the spatial information about the wetland landscape. Patch richness (PR) is important components of the landscape structure considering the way that the landscape components present in a landscape can impact an assortment of natural processes. Patch density (PD) decrease when the wetland fragmentation occurs frequently. The higher value of Largest patch index (LPI) denotes greater impact for maintain environmental process. Shannon diversity index (SHDI) helped to understand the wetland diversity in a specific area. SHDI value of ‘0’ indicates that in the study area, only one patch class exists and has no decent variety. The above indices were calculated by using Arc GIS 10.3.1.
Apart from the state indicator, we have selected two response indicators i.e., wetland degradation rate (WDR) and ecosystem service value (ESV). WDR is determined how much wetlands area has decreased within a particular period of time. Based on the previous research paper (Xie et al. 2001) we calculated ESV of the wetland classes in the study area.
2.2.5. Calculation of indicator weights and assessment methods
Saaty’s AHP technique is used for the determination of weight of the indicators (Saaty, 2008). The AHP priority was calculated via the website (https://bpmsg.com), which is freely accessible. Table 3 shows the correlation matrix and weight of each indicator used in AHP technique. For the construction of WEHI, the weight of the indicators is multiplied by respective standardized value of the indicators.
where, WEHI is the wetland ecosystem health index, Wi is the weight of the ith indicator, and Ci is the standardized value of the ith indicator.