Evaluating the Impact of Different Vegetation Types on NEE: A Case Study of Banni Grasslands, India

Estimation of NEE of Grasslands ecosystems becomes mandatory as these grasslands with their wide spread (almost 40% of land of the earth) and high plant diversity play a major role in global carbon balances and NEE at both local and global scale. The present study has been focused on understanding the role of different plant species responsible for variation in NEE of the Banni Grasslands of India. These grasslands form a belt of arid grassland having low growing forbs, graminoids and scattered tree cover. Due to its wide spread and inaccessibility of Banni, this study utilized spatial approach for evaluating carbon emissions and NEE. Landsat data was utilized for vegetation type classification and SMAP data for extraction of NEE values proved their potential for categorising vegetation type and generating NEE values precisely. Three major plant types were identified from the study area viz., Grasslands, Land with Acacia and Land with Prosopis. Grasses were dominant covering 77% and the rest of the area was occupied by the other two classes, i.e. Acacia and Prosopis. The NEE values were higher for the grasses when compared to the other two plant species proving to be the active sinks when compared to other plants. The differential contribution of NEE by species has been depicted in the present work.


Introduction
Regional and interannual patterns of the terrestrial carbon dynamics are chang- that depend upon the tower height, canopy physical characteristics and wind velocity [24]. This has unveiled the limitations of this technique despite its high temporal resolution [25]. Chamber technique is important for point measurements of NEE, but prone to a variety of potential errors and consumes plenty of time [26] [27] [28]. In addition, it is not possible to scale up net CO 2 exchange over the ecosystem level using EC and Chamber techniques [29] [30] [31]. Application of spatial approach can overcome this problem. Spatial approach also aids in understanding the distribution of different plant species of grasslands which helps in precise understanding.
Distribution of different grasses, their association and net primary production of these ecosystems play crucial role in global carbon budget and influence regional climate by modulating the evapotranspiration flux [13]. In this context, distribution and quantification of vegetation become imperative for understanding to encompass the changes in carbon flux. Satellite-based Normalized Differential Vegetation Index (NDVI) helps in comprehensive monitoring and quantification of vegetation [32] [33] [34]. Fractional Vegetation Cover (FVC) which is derived from NDVI using empirical relations is the vegetation-covered fraction of ground [35]. FVC is the ratio of the vertical projection area of vegetation (including leaves, stalks, and branches) on the ground to the total vegetation area which is directly detectable by the sensor from any view direction [36] [37].
FVC measured using spatial approach provides basic data for characterizing ecosystems which plays an extremely crucial role in the study of regional ecosystems [38] [39] [40] [41] [42]. Most importantly, FVC helps in understanding the seasonal changes occurring in the exchange of CO 2 between the land surfaces and the atmospheric boundary level [43].
Considering above facts, this study has been taken up to analyse the role of different plant species in variations of NEE to understand the carbon dynamics of Banni grasslands. These ecosystems are very sensitive to future changes in climate, and understanding how these systems have responded to climatic changes in the past can provide us with insights into their potential responses to future global change.  Traditionally, the Banni was declared as a Rakhal (reserve grassland) where only milch cattle were allowed to graze, and sheep and goats were not allowed to reduce the pressure on the grasslands. People were not allowed to reside in the Banni. Later, sheep and goats were also allowed to graze in the area but grazing was regulated by imposing fee at various rates for different categories of livestock. However, this traditional resource management system which had helped in the maintenance of equilibrium between environmental system and human activity since several centuries was no more functional [44]. The grazing regulations slowly disappeared, and all kinds of livestock from every part of the state and neighbouring states were allowed into the area. Large numbers of livestock used to immigrate for grazing during 3 -4 months of monsoon [45] [46]. Recent interventions such as introduction of P. juliflora, introduction of additional livestock have led to reduction in carrying capacity of these grasslands.

Pre-Processing the Datasets
Cloud-free Landsat ETM+ satellite data of Nov. 2017 was acquired from USGS website. The data was geographically corrected and was having geographical projection with WGS 84 datum. The study area was included in two different scenes. These two scenes were stacked separately and then mosaicked together to get the study area. The dataset was having the spatial resolution of 30 m × 30 m.
Daily Soil Moisture Active Passive (SMAP) version 4.0 NEE data from the month of January to December 2017 i.e. for 365 days was acquired from the website (https://nsidc.org/data/smap/data_versions). This data set was having the spatial resolution of 9 km × 9 km. The dataset was having geographical projection with WGS 84 datum. This data was subset for the study area and then stacked month wise for further processing.
where NIR denotes the near infrared band and Red denotes the red band.
FVC of the study area was derived using NDVI image using the following formula: where NDVI denotes the NDVI value of the pixel, NDVIveg is the NDVI value of a pure green vegetation pixel, and NDVIsoil is the NDVI value of bare soil.

Results and Discussion
Grasslands are amongst the important ecosystems that sequester and store large amounts of soil carbon, which is highly dependent on the factors like herbivory        as it may be assumed NEP to equal-NEE [61]. Positive values of the NEE indicated that the ecosystem was acting as a source during these months [62] [63].
The reason being in summer, water deficits caused leaf senescence (herbs) and therefore less assimilation of carbon leading to decrease in Gross Primary Production (GPP) and thus positive NEE. Periods of negative NEE (indicating net ecosystem uptake) were smaller in magnitude and spanned a shorter duration and coincided with the growing period of the grasses (monsoon). Acacia ecosystem varied from a net sink to a carbon source depending on the time of year, with a lower/higher magnitude during the warm/cold season [64].
Yearly comparison of the NDVI, FVC and NEE showed that the NDVI and        year 2017 as compared to the years 2015 and 2016. This indicated that the Ecosystem CO 2 exchange rates were strongly influenced by the type of vegetation which was also evident from the NEE values obtained for these categories. NDVI of the Banni grasslands was found to be higher than a threshold of about 0.3 which indicated that grasses were strong enough to drive a substantial portion of the NEE flux and provided improved NEE in comparison to the Prosopis and Acacia. For each category a steady increase with the growing season with a distinct decrease after the peak season was noted. With respect to the NEE values for different vegetational categories, not much difference was observed. Though Prosopis showed extreme higher and lower values ranging from −1.75 µg/m 2 to 0.73 µg/m 2 the cumulative values for NEE was highest for Grasses i.e. −0.5 followed by Acacia and Prosopis; 0.18 and 0.0, respectively proving grasslands to be an effective sink of carbon dioxide sequestration and no contribution of Prosopis for the same (Figure 17). Prosopis showed more negative values of NEE as compared to Acacia during the growth period while it showed more positive values during the summers. This indicated that though Prosopis absorbed a high amount of carbon and acted as a sink during growth period but at the same time released a high amount of carbon and acted as the source of the carbon. This may lead to a decreased capacity of carbon uptake and conversion of this ecosystem into the source of the carbon.

Conclusion
This study proved that NEE varies with different plant species and in this study, grasses were found to be the most active sink of the carbon dioxide while Prosopis and Acacia acted as weaker sinks. The ecosystem of Banni acted as the source of C for almost an entire year in the absence of grasses (i.e. during January to June) while as a sink during the growth period of grasses (i.e. during July to December). Furthermore, application of spatial approach provided consistent and systematic observations for monitoring plant species of Banni grasslands and NEE data at fine temporal resolutions which helped in comprehensive understanding.