Assessment of Carbon Sequestration and Sink Potential: A Surrogate for Ecosystem Carrying Capacity in Dolvi Region, India

: The socio-economic development activities are aimed to improve the quality of life of the inhabitants along with industrial growth of the region. The development activity proposing expansion or new construction is anticipated to affect environment and hence any ecosystem existing in the vicinity of development activity needs to be assessed. In this study, an attempt has been made to assess the changes in existing ecosystem services due to expansion of industry and construction activities in the Dolvi region. For assessing the carrying capacity of region in such scenario indicators like NDVI, NDWI, Vegetation NDWI and forest ecosystem services such as carbon stock are estimated using geospatial techniques. To account for the possibility and effect of development on forest ecosystem, remote sensing and field analysis was carried out to estimate the status of indicators along with carbon sequestration and sink potential for the region. The NDVI value of 0.87 show healthy vegetation existing in the region. The NDWI has indicated the presence of water indicating region is not under water stress. From the current biomass estimation


Introduction
The estimation of assimilative capacity of biological environment or in an ecosystem is the capacity to adsorb or absorb pollutants without any damage to flora.In nature, Carbon sequestration is the process that absorbs carbon dioxide (CO 2 ) from atmosphere and store it in floral ecosystem.Photosynthetic plants and trees act as a sink for CO 2 and with more photosynthetic activity more CO 2 is converted into biomass and sequestering it in all parts of plants including leaves, stems, branches, trunk and roots in the form of carbon [2] Maintaining these Carbon stocks are also important to enhance related ecosystem services [10] due to linkage between soil organic carbon stock and atmospheric concentration of CO 2 [17].The estimation of carbon sequestered in any ecosystem is important as growing anthropogenic activities increases the emissions and an increased carbon sequestration can help regions to maintain its ambient atmosphere by sequestering large amount of Carbon and increasing its carbon stock in form of biomass [5].The estimation of carbon stock in any ecosystem helps in keeping check on deforestation with a sustainable forest management can help policy makers for preservation and conservation [16].The enhanced carbon stock in any ecosystem ensures the human social development and hence knowing the stocks and sinks becomes very important [19].
The estimation of total biomass and carbon required extensive field survey and laboratory analysis.Allometric measurements are required for multiple plants at multiple A Surrogate for Ecosystem Carrying Capacity in Dolvi Region, India locations to predict biomass and carbon accumulation [1].This estimation process is labour intensive and takes longer to estimate total biomass.Advance remote sensing techniques combined with field data have shown promising results.The high-resolution data of images captured through sensors has given appreciable results for estimating crown area or crown diameter in tropical forest trees [3].Remote sensing methods are highly preferred due to its faster estimation and ability to extrapolate forest biomass compared to LIDAR.The analysis using data obtained through this has shown positive and higher correlations between sensor metrics and different measurements of forest structure [12].The objective of the study is to assess the changes in existing ecosystem services due to expansion of developmental activities in Dolvi region of India.

Study Area
In order to, ensure that there is no further degradation of the environment because of expansion plans and development it becomes necessary to assess the ecosystem carrying capacity of Dolvi region.The remote sensing-based studies provide knowledge of spatial distribution of natural resources.The satellite images and indices imply that within 10 km.radius, there are rich and different plant ecosystems like terrestrial dense forests, agriculture and mangrove plants and aquatic ecosystem.These ecosystems provide shelter to other diverse types of flora and fauna at the region.An area of 10 km radius around the industrial development site, depending on the important topographical attributes such as land use, water body, drainage, eco-sensitive areas and locations of habitats, trunk key constructions including roads, railways, pipelines and industries if any in the area are to be mentioned.Proposed project site and study area are located at Latitude: 18°41'13.77"North and Longitude: 73° 1'46.78"East.Radius from plant site (for Biodiversity assessment), Surrounding area of 1km (core zone) and 10 km (buffer zone).

Methodology
Carbon sequestration and sink potential as a surrogate of carrying capacity for a region with different ecosystems helps in identifying the potential of region for its long-term sustainability.The vegetation health and availability of fresh water are crucial for thriving species richness in an area.
The methodology for finding the Carbon sequestration or Carbon stored or Storage potential of region was found using the methodology shown in Figure 2.

Data Collection
For studying the terrestrial ecosystems using remote sensing, the vegetation and water indices are the most important tools of estimating and comparing ecosystem health For this Sentinel-2 satellite image of January, 2020 was used for estimation of indices in the study.Also to estimate the health of surrounding vegetation and its distribution for estimation of sink potential of the area.For estimation of biomass vegetation structural information like canopy height, canopy cover, canopy density etc. are useful properties.These properties cannot be estimated using conventional optical remote sensing.Active remote sensing techniques like "Synthetic Aperture Radar (SAR)" and "Light Detection and Ranging (LiDAR)" have proven extremely useful methods.The recent remote sensing data provided by NASA's Global Ecosystem Dynamics Investigation (GEDI) mission is provides LiDAR data for global forest covers.The mission involves powerful LiDAR instrument attached to the International Space Station (ISS) which started providing data from April 2019.Although the objective of mission is estimating ecosystem health and biomass, currently only raw data with canopy parameters are available.The GEDI data was used to estimate average tree height at some of the regions where sampling was inaccessible.

Forest Height (FH) Data
GEDI mission data is a point data providing different canopy characteristics based on LASER pings.It provides precise information about canopy height, canopy vertical structure and ground elevation.Although this information is very crucial for biomass and structural information, the point wise data is available linearly throughout the study area.Multiple datasets were downloaded and processed using Rstatistical package.The "rhdf5" package was used to process HDF5 data format and extract height information.The height data was rasterized by interpolation to obtain continuous data for analysis.

Forest Canopy Height
It is the height of the vegetation that has its highest components above ground level, it is important for micrometeorological parameters and is required in estimating forest biomass and carbon pools.

Soil Adjusted Vegetation Index (SAVI)
Another important vegetation index, Soil Adjusted Vegetation Index (SAVI) was used to provide accurate measures of vegetation removing any effect of soil.This gives better separation between soil and tree canopy.L is usually taken as 0.5.

Indicators and Analysis
The indicators to assess the health of floral species in an ecosystem are Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI) and Vegetation NDWI.These are selected as indicators since it not only assesses the current status of the ecosystems, but can also be compared to future observations.NDVI values are significantly associated with human health conditions.Green spaces around residential locations are good for air and noise pollution reduction and natural healing (Jason et al, 2019).

Normalized Difference Vegetation Index (NDVI)
To assess the floral ecosystem using remote sensing data with spectral indices NDVI was used.Using NDVI, chlorophyll present in the plant ecosystem was estimated.NDVI was estimated by taking the ratio of difference and addition of NIR (Band 8) band and Red (Band 4) bands of remote sensing image.
(2) Sentinel-2 satellite's Multispectral Instrument (MSI) to obtain the remote sensing images at Dolvi region.An image obtained in January 2020 was used to avoid interference of clouds and dust in the images.The MSI sensor provides 12 bit optical remote sensing data at 10m spatial resolution.Due to high spatial resolution, the sensor is highly preferred for vegetation and terrestrial remote sensing.

Normalized Difference Water Index (NDWI)
Fresh water reflects the blue-green wavelengths while absorbing the NIR wavelengths.This information was used A Surrogate for Ecosystem Carrying Capacity in Dolvi Region, India to derive the NDWI.NDWI was derived by using Green (Band 3) and NIR (Band 8) bands of the image. (3)

Vegetation NDWI
Vegetation NDWI was estimated for water stressed detection of the vegetation by assessing water availability in the leaves of green plants.Vegetation NDWI was estimated using near infrared (band 8A) and shortwave infra-red (SWIR) and (band 11) of Sentinel-2 data respectively.Both bands were available at 20-meter resolution for Dolvi region.
Once the health of floral ecosystem is identified, the estimation of above ground biomass (AGB) and total biomass was done to identify the carbon sink potential of the region.For estimating the AGB different species of trees were selected, and non-destructive method was used and using diameter at breast height (DBH), AGB was estimated.

Diameter at Breast Height (DBH)
DBH is the diameter of tree measured at 4.5 feet above ground, and is measured with a specially calibrated diameter tape, a d-tape that reflects the diameter measured when put around the circumference of a tree.The girthing or d-tape with calipers are the most common instruments for DBH measurement [2] and circumference of trees were measured.DBH estimates the timber volume of a single tree by using the allometric correlations between stem diameter, tree height and timber volume.

Girth at Brest Height (GBH)
In this girth of tree was measured at approximately 1.3 m above the ground and the GBH of trees with diameter more than 10 cm were measured directly by tape.Trees with a diameter of more than 10 cm were considered for biomass assessment [6,9,14,18].
DBH is diameter at breast height (cm).

Total Biomass Estimation
The biomass of a tree is the sum of the biomass of its trunk, root, branches, leaves and other organs like fruits and flowers etc.For an accurate estimate of biomass, a tree would have to be felled, burned and weighed to get dry biomass and carbon.AGB in this study was estimated using tree's bio volume.Volume estimation was done by finding the Basal Area (BA) of a tree.BA is the cross-sectional area at the base of the tree.It is closely related to volume and is useful measure of ground occupancy by the tree.It is given by the following formula: where, BA is Basal Area (in cm) The natural vegetation does not follow geometric shape in general.Usually tall straight trees are conical at top and have wide basal growth.Hence tree volume is decided by an empirical formula involving a form factor to compensate tree shape.
Where V is Tree Volume in m 3  BA is Basal Area in m 2 H is Height of tree in m The calculated volume has been used to estimate wood biomass.Since the wood density varies with region and tree species, mean density value of 0.57 gm/cm 3 estimated for tropical Asian region was used.
Total Biomass = Specific gravity of wood × Volume

Carbon Sequestration and Storage Potential
Generally, 50% of biomass for any plant species is considered as carbon [15].Therefore, the weight of carbon in the tree is calculated by multiplying the biomass of the tree by 50% [8].

Results and Discussion
The indicators used as a surrogate to estimate the carrying capacity of ecosystem in Dolvi were estimated and analyzed using GIS and remote sensing data and thus the carbon sequestered and sink potential was estimated.

NDVI
It is based on spectral reflectance of the Chlorophyll molecule present in green vegetation.The chlorophyll present in leaves reflects green and near infra-red (NIR) wavelengths from leaf surface while absorbing the red wavelengths [7].This creates a signature spectral response curve for the chlorophyll molecule separating it from other surface objects present in remote sensing image.NDVI is used as a proxy parameter for vegetation productivity.Absorbed radiation through photosynthesis is estimated along with other features such as leaf area index, vegetation biomass, and water availability.
Fresh water on the other hand reflects the blue-green wavelengths while absorbing the NIR wavelengths.This information helps derive the Normalized Difference Water Index (NDWI) [11,13].Using these properties, both vegetation and water can be separated in remote sensing images.
To estimate the NDVI, by avoiding the effect of atmospheric scattering and absorption, an atmospherically corrected surface reflectance image was used.Google Earth Engine was used to process the imagery.The images were processed in QGIS and the obtained output was in a normalized range of values from -1 to 1. Values below zero indicated absence of any vegetation at that location.Values close to -0.5 were usually water or built-up area present on the ground.The positive values indicated presence of vegetation on the ground.Very high values close to 1 indicated healthy green vegetation or green agriculture crops.The obtained NDVI values for the region are shown in Figure 3.

NDWI
NDWI Map as shown in Figure 4 also gives values ranging from -1 to 1. Ideally values below 0.3 are non-water while values above 0.3 are water bodies.The NDWI is used to identify existing water bodies in the remote sensing data and monitor the water content of the region.
The Figure 4 shows presence of multiple freshwater bodies apart from the main river.This water availability is beneficial for human settlements as well as regional wildlife present.Ground water recharged by such water bodies is available to the forests and shrub vegetation.

Overlay Analysis of the Raster Layers
The overlay analysis of NDVI, Carbon, NDWI and Topography layer shown in Figure 6 suggests that the topography of the region plays important role in harboring the biodiversity of the region.Undulating topography helps create surface runoff and natural water bodies required for local flora and fauna.High NDVI values are observed at high elevations present within the 10km radius.This suggests the hill slopes provide better support to terrestrial ecosystems.These hill slopes can be used to improve the vegetation by carrying out plantation at selected locations as natural vegetation is observed at these locations.

Biomass Estimation and Sink Potential
In this study, the analysis of satellite images to estimate the seasonal dynamics of carbon sequestration was undertaken.Optical remote sensing has been used to estimate different vegetation properties based on spectral reflectance curve.The maximum information about vegetation health is provided by Red and NIR bands of optical remote sensing data.The optical remote sensing techniques revolve around chlorophyll properties.

Biomass Estimation
The biomass information was modelled with NDVI, SAVI, FH and DEM using linear model function (lm) to check linear relationship.A lower value of R 2 (coefficient of determination) as 0.42 was obtained.The data was split into 70:30 ratio as testing and training data.70% data was used to train 'Random Forest' (RF) model.Remaining 30% was used to test the model.Based on this information the RF regression model was built using R statistical package.The model was trained using "mtry" and "tuneGrid" parameters to optimize the analysis.The model did not show very improved performance as compared to the linear model.The obtained R 2 value was 0.44 showing slight improvement.The tuned model was later applied for prediction application.The raster layers of NDVI, SAVI, FH and DEM were inputs for the model.The output model is a map of biomass distribution with unit of Mg/ ha.The model uncertainty was estimated as pixel wise probability.
The predicted biomass for overall distribution of biomass at each pixel location is shown in The existing biomass estimated at the site location is the biomass of living woody trees which sequester carbon through photosynthesis.From the current biomass estimation, the steel Plant and 10 km region shows presence of forest which has biomass as high as 8 Mg/ha at terrestrial vegetation and 5 Mg/ha at mangrove vegetation.From the current estimate it can be concluded that the region supports living forest biomass which will continue carbon sequestration through photosynthesis.Hence the region has good carbon sequestration potential.The carbon sequestration potential and rate of sequestration changes need to be monitored over time depending on afforestation and deforestation activities.Planting more trees will ensure increase in the carbon sequestration potential of the region.Protecting the existing forest in addition to afforestation activities can increase the carbon sequestration potential.

Carbon Sequestration
Carbon sequestration is the storage of atmospheric CO 2 as solid carbon in plants for long-term, soils and ocean and geologic formation through natural or anthropogenic activities.Carbon sinks are reservoirs that store carbon and thus maintaining Earth's atmosphere.Carbon sink potential is the potential of sinks to sequester amount of carbon per year.The carbon accumulation at the region is computed with the factor of 0.48 (Figure 8) Similar to vegetation high carbon values are present at the elevated forest regions and some mangrove regions.This shows the natural forest vegetation has high potential for carbon sequestration.Area near steel plant shows low values.These are result of some vegetation which is present and estimation error.At north of the 10km radius, some high carbon values are present near the agriculture fields.Mangroves on both sides of the river also have potential to sequester atmospheric carbon.To identify the precise areas of high carbon sequestration, the area was divided into grids of 5km x 5km size.Each pixel of the map is 300m x 300m in dimension.The carbon values of the grids are derived from the above ground biomass.Since the grids contain different land cover classes, averaging the values over the grids will not provide a justified result.Hence maximum carbon values observed in the grid was estimated.
Grids 14, 15, 19, 20, 24 and 25 show highest amount of carbon sequestered in form of terrestrial forests and grid 4 show least amount of carbon values present.Grids 1, 2, 6 and 7 show carbon values due to the presence of mangrove vegetation.These grids have high scope for the improvement of mangrove biomass by carrying out mangrove plantation.Similarly, vegetation in grids 13, 16, 17 and 18 can provide land area for afforestation activities.
The grid wise analysis of NDVI also implies similar trends.However, mangrove grids 1, 2, 6 and 7 are equally important even though NDVI values are not very high.The current steel plant also lies in grid 13.To curb air and noise pollution, vegetation can be increased at low NDVI and low carbon grids.

Conclusions
The estimations for ecosystem carrying capacity have been carried out to identify the Carbon Sink Potential and Carbon sequestered through indicators i.e.NDVI, NDWI, Vegetation NDWI, Soil Adjusted Vegetation Index (SAVI) etc.The sink potential and carbon sequestered helps in identifying the potential of region for its assimilative as well as supportive capacity.But the carbon relationships are scale dependent and hence sink potential of the region serves as a requisite for conservation of biodiversity.
This study reveals that NDVI values in the region maximum were at 0.87 indicating healthy vegetation in the region and most vegetation was present on undulating terrain or as mangroves.The vegetation NDWI was used for water stress detection and the results of this analysis represents the availability of water in leaves of green plants.The value of this indicator above 0.3 shows healthy vegetation with sufficient water content.The biomass analysis showed lower values from 0.8 to 5 Megagrams/ha at low flat regions while mangrove biomass valued around 9 to 10 Megagrams/ha and high values of biomass was observed at elevations where dense vegetation was present and was around 20 Mg/has the region has high biomass content, it will continue carbon sequestration through photosynthesis and hence a region with good carbon sequestration potential and any activity of afforestation or deforestation will change the region's sink potential.It was also concluded that region has good Carbon stock acting as a sink for absorbing emissions from the construction and developmental activities in the region.

Recommendations
In this study, the role of vegetation in mitigating the suspended dust was discussed and assessed through remote sensing.The sink potential of the region for air pollution can be enhanced and validated if regular monitoring of vegetation stock is done in parallel to emissions in the region.This will help in establishing the effect of emissions on sustained growth of vegetation as they are the first indicator of any disturbance in the region.Temporal variation of the region for its carbon stock and relation with emissions can also be done in the future study.

Figure 1 .
Figure 1.Details of Study area with different Ecosystem Services.

Figure 2 .
Figure 2. Methodology used in this study.

Figure 3 .
Figure 3. NDVI map of study region.

Figure 3
Figure 3 map shows Amba River (dark red) flowing from south-east to north-west at the region showing presence of mangrove vegetation at the banks.The NDVI map shown here indicates regions of dense vegetation away from the riverbanks.Large mangrove vegetation is present on the south-west side of the riverbank at middle.The NDVI values in the region ranged from -0.46 to 0.87 indicating presence of healthy vegetation for values between 0.6 to 0.9 and moderate NDVI values for range 0.2 to 0.5 values and low NDVI for less than equal to 0.1.Most of this vegetation was present on the undulating terrain or as mangroves.Since the data was collected during pre-monsoon season, agriculture land did not show any crop vegetation.

Figure 4 .
Figure 4. NDWI showing fresh water bodies.Both the maps indicate overall healthy vegetation without water stress and availability of freshwater bodies which can support local fauna.

Figure 5
Figure 5 reveals vegetation NDWI for water stress detection.The output shows water availability in the leaves of green plants.NIR reflectance is affected by dry biomass than water content of leaf tissue.Whereas, SWIR reflectance depends upon vegetation water and mesophyll present in the leaves.Hence normalizing their values provides water stress in the leaves.

Figure 5 .
Figure 5. Map of vegetation NDWI showing water stress at the location.The vegetation NDWI values above 0.3 shows healthy vegetation with sufficient water content.The blue regions at terrestrial and mangrove region indicate the vegetation is not

Figure 6 .
Figure 6.Overlay analysis of the raster layers.

Figure 7 .
The map shows presence of low biomass values from 0.8 to 5 Mg/ha at low flat regions near steel plant.The mangrove regions nearby show moderate biomass values around 9-10 Megagram/ha.High values of biomass was observed at elevations where dense vegetation was present.The regions show biomass values close to 20 Megagram/ ha at some locations.

Figure 7 .
Figure 7. Above Ground Biomass predicted using remote sensing.