Measuring, forest cover of which 23 km2 are managed

monitoring and evaluating the estimation of biomass is the core foundation of
REDD+ program. So, the aim of this study is to calculate the AGB in Kayer Khola
watershed (Chitwan District) which is one of the pilot project site in Nepal
(Bajracharya, 2011). Extensive field surveys, sample plot collection and
researches already been done in this area. The availability of researches like
use of airborne LiDAR with ground survey data (Karna, 2015; Mbaabu, 2014) or integrated use of Geo-Eye and
ALOS PALSAR data (Qazi, 2017)
for deriving AGB estimation would be beneficial. The major objectives of this
study is the use of SAR data for the biomass estimation using methodology like
data processing, statistical regression and algorithms for betterment of
results. Moreover, since the sample plots is being used as that of the earlier
study (Gilani H, 2015) a comparison
could also be made.  Study Area The Kayarkhola
Watershed is located in the Chitwan district of the Central Development Region
of Nepal. Altitude of this watershed ranges from 245 m to 1944 m, covering area
of 8002 ha. Out of the total area 5821 ha area is forest cover of which 23 km2
are managed by 16 community forest user groups and 74 leasehold forests cover
area 248 ha. The study area comprises of lower tropical to lower subtropical
forests, bounded by 27.6680N to 27.7760N latitude and
84.5560E to 84.6950E longitude. The region lies in the
tropical monsoon climatic zone with high humidity throughout the year, with
average annual rainfall of 1510 mm/year.  Datasets Forest InventoryForest field data
were collected from 51 circular survey plots of radius 12.62m each for the
study. Tree parameters like diameter at breast height only equal or greater
than 5 cm, height only equal and greater than 2 cm, canopy cover, species and
crown diameter were measured in these plots. The Field survey were carried out
from December 2009 to December 2010 for plot level biomass estimation. (Qazi, 2017) ALOS PALSAR MosaicThe ALOS PALSAR
25m mosaic data was downloaded from years used in
this were 2009-2010 and 2015-2016. We couldn’t use data from 2011-2014 as the
mosaic weren’t available. Data earlier to 2009 haven’t been used due to the
field data time period. SentinelSentinel 1 data,
level-1 and GRD product with VV and VH polarization has been downloaded from This acquisition
date of this data is 2015/04/29.  ALOS PALSAR-2ALOS PALSAR 2 data
has been purchased with the acquisition date being 2014/11/01. The data is a
quad polarization with 3.12 m of pixel spacing.  Research FrameworkThere
are various methodologies for biomass estimation but no current methodology presents
a clear view on how carbon pools and their fluxes should be reported and what
the accuracy and uncertainty of biomass monitoring might be. (Avtar, 2013) Therefore the need to map biomass and
produce map data on the forest carbon stocks and changes are essential. The
traditional way of destructing the forest and calculating the biomass is not
just time consuming but negating the whole idea of biomass estimation. Hence
the remote sensing data supplemented with forest inventory data can provide
cheap and fast estimation as well as historical information about forest
biomass. Most of the remote sensing techniques are based on optical and
synthetic aperture radar systems. These techniques could have serve as valuable
methods for biomass assessment of heterogeneous complex biophysical
environments. SAR has been effectively used for assessing forest biomass
through the campaigns of airborne and space borne SARs. Special enhances have
been lead to Polarimetric SAR (PolSAR) which could be a suitable alternative
with active development particularly in forestry applications (Antropov, 2017).  FIELD DATASquare
plots design used to facilitate the pixel sampling based on satellite data to
reduce position error. A systematic random sampling design was applied for the
purpose and forest inventory parameters like tree height, species, tree
density, etc. were measured. Sampling plots mostly selected in plain area to
minimize topographic effects of SAR data (Avtar, 2013).
Rectangular plot of field measurements with subplots were collected by the NASA
DESDynl. The biomass of tree and small stems were separated by different
diameter category. NASA’S Laser Vegetation Imaging Sensor (LVIS) is an airborne
laser altimeter system, LVIS Ground Elevation (LGE) data were used in the study
which included location, ground surface elevation and heights of energy
quartiles. They were a form of relatively direct measure of the vertical
profile of canopy components. (Ni, 2014).
Random stratified sampling techniques was used for square field plots and
efforts were made to select samples to include maximum variability of available
forest age groups in homogeneous forest regions with minimum visible variation
of structure and composition (Baig, 2017). Stratified random sampling was
used to determine the optimum number of sample plots and data collection was
used for tree crown delineation, tree species classification and validation (Mbaabu, 2014).  RASTER DATAPALSAR
FBD 50m mosaic data with HH an HV polarization. The processing of data was
started with the terrain correction to minimize the topographic effects in
mountainous areas using SRTM DEM 90m. The DEM has been resampled to 50m to
align with PALSAR mosaic data. (Avtar, 2013). A level
1.1 PALSAR data have been used in the study with dual polarization mode. The
ASTER GDEM was transformed to ellipsoid surface which was the vertical
reference of digital elevation model derived from PALSAR InSAR data. (Ni, 2014).  Sentinel-1 VV-polarized, ALOS-2 PALSAR-2
HV-polarized and Landsat (ETM and OLI) used (Reiche, 2017).
SAR (ALOS PALSAR-2) dual polarization imagery used for AGB estimation and high
resolution optical (WorldView-3) data used for visual interpretation and ground
survey (Baig, 2017). The global dual
polarization (HH, HV) 25m resolution ALOS PALSAR mosaic for mapping the African
continent, about 180 image strips used (Bouvet, 2018). Two types of
remotely acquired data like airborne LiDAR data and Geo-Eye-1 satellite images
are used for their study (Mbaabu, 2014). A “benchmark” map of biomass
carbon stocks over 2.5 billion ha of forests on three continents has been
mapped using a combination of data like Lidar, optical and microwave imagery to
extrapolate over the landscape (Saatchi, 2011). PolSAR scenes
with reference data like that of forest inventory has been used (Antropov, 2017). The ALOS PALSAR dual polarization data
were processed including terrain correction, radiometric calibration and
backscatter coefficient derived for forest biomass estimation in western terai
region of Nepal (Nashrrullah, 2012) METHODSThe
penetration depth was obtained from the difference between Digital Scattering
Phase Center Model and Registered ASTER GDEM. The DSPCM can be produced after
the atmospheric effect was removed from InSAR phase (Ni, 2014).
Pre-processing, extraction of NDVI time series from Landsat and co-registration
of Sentinel, ALOS PALSAR data to Landsat NDVI using GAMMA software (Reiche, 2017). In SAR imagery, salt and pepper like
texture known as speckle noise degrades the image quality and makes feature
interpretation difficult. GAMMA Map spatial filter with 3X3 window size was
applied over the calibrated SAR data (Baig, 2017). To find the relationship between
radar backscattering and AGB depends on a large number of parameters that are
related to vegetation or ground or collection environment. So, the approach
consists in performing the AGB estimation separately in the “wet season” and in
the “dry season” by adding conditions like leaf-on, wet soil and high vegetation
to the “wet season” and leaf-off, dry soil and lower vegetation water content
to “dry season” (Bouvet, 2018)  STASTISTICAL ANALYSISMulti-linear
regression (MLR) analysis conducted to relate the backscattering of PALSAR to
filed calculated biomass. The size of the sampling window was 3X3 pixels.
Finally validation was used to evaluate the accuracy of the model by comparing
PALSAR estimated AGB to the field derived AGB. The MLR model for biomass
estimation was developed using backscattering HV and Backscattering HH/HV
because HV & HH/HV shows strong correlation with biomass. (Avtar, 2013). Data using probabilistic approach
combining Sentinel, ALOS PALSAR and Landsat NDVI time series to detect
deforestation in NRT(capacity to detect new changes in satellite images once
they are available) during the monitoring period. Using spatial normalization
to reduce dry forest seasonality in the data (Reiche, 2017).
General form of regression used for the correlation study between the AGB and
the backscattering coefficients. Due to small number of total sample plots the
leave one cut cross validation jackknife regression method used for model
validation applied to HH and HV polarization (Baig, 2017). A data fusion model based on the
maximum entropy approach was applied. The model produced a map of AGB along with
estimates of uncertainty at a spatial resolution (Saatchi, 2011). In their study
they have used a non-parametric estimation technique: k nearest neighbor (kNN)
regression method this was used for reducing the peculiarities like solving two
non-linear equations (Antropov, 2017).
Pearson’s correlation between the backscatter and the AGB was calculated for
different forest type.  RESULTSThe
backscattering HV polarization produces better correlation than backscattering
HH because of the volume scattering in forest areas enhances the
cross-polarization returns with the increase in biomass. Backscattering HV is
less influenced by soil and vegetation moisture than backscattering HH. A loss in
sensitivity of PALSAR signal appeared to occur at approximately 150-200 Mg/ha
biomass. (Avtar, 2013) The results showed that
introduction of penetration depth could clearly improve the accuracy of forest
biomass estimation. The correlation between the predicted biomass and the field
measurement over sampling plots were improved (Ni, 2014).
Deforestation events were detected with higher spatial and temporal accuracies
when combining observations from multiple sensors than when using observations
from a single sensor. Combining multiple SAR and optical time series can
guarantee regular and temporally dense observations at medium spatial
resolution independent of weather, season and spatial location which can
improve NRT deforestation monitoring in the tropics (Reiche, 2017).
The benchmark map provides a spatially refined and methodologically comparable
carbon stock estimate for forests across 75 developing countries and improves
upon previous assessments based on often old and incomplete national forest
inventory data (Saatchi, 2011)                                        SUGGESTIONS

The saturation
problems of SAR data can be overcome using Polarimetric-interferometry SAR
(PolInSAR) technique or P-band SAR data. For more precise we must look forward
for the P-band or DESDynl satellite system in the future. (Avtar, 2013). Forest spatial homogeneity was another
important aspect which should be considered in the forest biomass estimation
using penetration depth. There are still many issues that need to be explored
in our future research before it becomes a practical method, including the
effects of forest disturbance the temporal decorrelation (Ni, 2014). More efforts should be made to remove
the terrain effects with an appropriate model-based slope correction.
Integration with other remote sensing data such as LiDAR or other very high
resolution may complement each other (Nashrrullah, 2012).

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now