Estimating Forest Tree Carbon using Remote Sensing Data and Techniques


Master's Thesis, 2013

79 Pages, Grade: 3.87


Excerpt

CONTENTS

LIST OF TABLES

LIST OF FIGURES

LIST OF NOTATIONS

LIST OF ABBREVIATIONS

Chapter 1
1.1 Background
1.1.1 Biomass estimation and remote sensing
1.2 Study Area
1.2.1 Location, demography and landuse/ landcover
1.2.2 Flora
1.2.3 Fauna
1.2.4 Livestock
1.2.5 Geology and soil
1.2.6 Tourism impact
1.3 Research Problem
1.4 Research Questions
1.5 Hypothesis
1.6 Research Objectives
1.7 Research Significance

Chapter 2
2.1 Remote Sensing Data for Carbon Estimation
2.2 Data Mining, GIS and Remote Sensing for Carbon Stock Assessment

Chapter 3
3.1 Activity Flowchart
3.2 Pre-Field Work
3.3 Sampling Design
3.4 Field Data Collection
3.4.1 Field data analysis
3.5 Aboveground Biomass/ Carbon Estimation from Field Data
3.6 Satellite Image Acquisition and Processing
3.6.1 Landuse/ landcover classification
3.7 Statistical Models for Carbon Estimation
3.7.1 Classification and regression trees (CART)
3.7.2 Random forests (RF)
3.7.3 Multiple regression analysis (MRA)

Chapter 4
4.1 Landuse/ landcover Classified Mapping
4.1.1 Area covered by landuse/ landcover classes
4.2 Field Inventory Data Analysis
4.2.1 Diameter at breast height (DBH) of trees
4.2.2 Height of trees
4.3 Field Biomass/ Carbon Stock Estimation per Plot
4.4 Biomass/ Carbon Stock Mapping based on Field Data
4.5 Parameters Extraction from Satellite Imagery for Statistical Analysis.
4.5.1 Normalized difference vegetation index (NDVI) extraction
4.5.2 Topographic factors extraction
4.6 Statistical Analysis and Carbon Stock Mapping
4.6.1 Simple linear regression analysis
4.6.2 Random forests (RF) analysis and carbon prediction mapping
4.6.3 Multiple regression analysis (MRA) and carbon prediction mapping
4.6.4 Classification and regression trees (CART) analysis

Chapter 5
5.1 Conclusions
5.2 Recommendations

REFERENCES

Annexure I

Annexure II

Annexure III

Annexure IV

Abstract

Alarming rates of carbon increase in the atmosphere threatens our planet for sustainability. For avoiding carbon release and sequester the carbon emissions from the forests and other sources, nations of the world are striving hard to design methodologies for monitoring carbon emission and sequester rates in the forests and other carbon pools. This book presents a research based on a very modern approach to estimate carbon in the moist temperate Himalayan forest of Ayubia National Park (ANP), Pakistan. Latest Geospatial techniques including Remote Sensing and Geographical Information System (GIS) were incorporated to estimate carbon in the forest of ANP using the up-to-date statistical approaches. Objectives of this research included landcover mapping of the study area and estimating and mapping the carbon in the forest as well as biomass for the same area. High spatial resolution Satellite Pour I’Observation de la Terre-5 (SPOT-5) satellite imagery is used to map the landcover of the study area. Different bands from Landsat-5 Thematic Mapper (TM) and SPOT-5 satellite images have been incorporated with inclusion of topographic factors to estimate the carbon stock. The results showed that three landcover types of ANP forest including mix forest, conifer forest and shadow conifer forest have 104,781.62, 445,290.61 and 301,972.37 tons of total aboveground carbon (AGC) respectively as calculated from field inventory data with an average of252.26 tC/Ha as compared to the previously calculated average of AGC of 222.99 tC/Ha at ANP. This research concluded that multiple and non-linear regression models can estimate carbon more accurately as compared to linear regression models. Linear regression correlations were poor while the Multiple Regression Analysis (MRA) method showed coefficient of correlation (R[2]) value 0.901 Random Forests (RF) prediction modeling showed R[2] value 0.90 while Classification And Regression Tree (CART) explained R[2] value 0.72. This proved that multiple and non-linear regression data mining models, such as MRA and RF, are best suitable while working with remote sensing techniques in natural systems like forests to calculate AGC.

LIST OF TABLES

Table 1.1: Landuse / landcover (LULC) classification of ANP (WWF Pakistan)

Table 2.1 Carbon pools identification by 150 countries (UNFCCC)

Table 2.2: Carbon pools estimation in ANP (Qasim, 2011)

Table 2.3: Summary of techniques and methods for above-ground biomass (Dengsheng, 2006)

Table 4.1: SPOT-5 satellite data characteristics

Table 4.2: Landuse/ Landcover area in hectare and percentage

Table 4.3: Estimated biomass per landcover in study area

Table 4.4: Estimated carbon per landcover in study area

Table 4.5: Topographic and satellite image parameters for carbon prediction

Table 4.6: Relationship of variables with target variable using RF

Table 4.7: Relationships of variable with target variable using MRA

Table 4.8: Sum of residual error for multiple regression analysis

LIST OF FIGURES

Figure 1.1: Biomass system generation and stages of CO2 emission

Figure 1.2: Study area map of Ayubia National Park (ANP) along with elevation

Figure 2.1: A conceptual approach for monitoring biomass in large scale (DeFries, et al., 2006)

Figure 3.1: Carbon mapping methodology flow chart

Figure 3.2: Field inventory survey map for taking sample plots

Figure 3.3: DBH is measured at 1.3 meters aboveground

Figure 3.4: Difference between circumference and diameter

Figure 3.5: Conifer trees in plot no. 5 in the left and on right surveyor measuring DBH of a conifer tree

Figure 4.1: Landuse/ landcover classified map of the study area

Figure 4.2: Area covered by each landuse/ landcover class in study area

Figure 4.3: Plot-wise distribution of trees with DBH in centimeter

Figure 4.4: Plot-wise distribution of trees with Heights in meters

Figure 4.5: Estimated carbon for each plot based on field inventory data of the study area

Figure 4.6: Aboveground Biomass (AGB) map of study area as per field inventory data calculation

Figure 4.7: Aboveground Carbon (AGC) map of the study area as per field inventory data calculation

Figure 4.8: Linear regression correlation of carbon with independent variables

Figure 4.9: Estimated aboveground carbon stock map of the study area using random forests prediction modeling

Figure 4.10: Residual plots for carbon

Figure 4.11: Estimated aboveground carbon stock map of the study area using multiple regression analysis

Figure 4.12: CART model findings

LIST OF NOTATIONS

Abbildung in dieser Leseprobe nicht enthalten

LIST OF ABBREVIATIONS

Abbildung in dieser Leseprobe nicht enthalten

CHAPTER 1

INTRODUCTION

Chapter 1
INTRODUCTION

1.1 Background

Environmental components are often being affected by the process of development as development often damage environment to some extent if started without considering carrying capacity. In order to avoid environmental problems, one must consider community, commercial and environmental aspects as these environmental problems generally arise as a result of economic growth. Out of these the major environmental concern is forest degradation and deforestation (Mya, 2010).

“The total amount of biological material (water removed) which is present above the surface of the soil in a specific area is termed as aboveground biomass (AGB) ” (Maharajan, 2012). 80% of terrestrial aboveground and 40% of terrestrial belowground biomass (BGB) carbon is being naturally stored by the forests on global scale as they participate in sinking carbon dioxide (CO2) intensities through sequestration of atmospheric carbon process of photosynthesis in plants and by aggregating soil organic content too (Baral, et al., 2009). One of the most common greenhouse gases (GHGs) is CO2 that put its part in rising global temperature which ultimately resulting in floods, draught, deforestation etc. Atmosphere, soils, forests and oceans are the major store houses of carbon and out of these “sinks" absorb more carbon than they release (Pareta & Pareta, 2011).

Global carbon balance is essentially controlled by the role of forests as they act both as carbon sources and sinks. Forests are one of the key elements to potentially combat global climate change and are natural ‘resistance’ to climate change process as they store and confiscate carbon more as compared to other global ecosystems on the earth (Holly, et al., 2007). In other words, forests perform a foremost part in keeping low the concentration of CO2 in the atmosphere (Mya, 2010). Climate change process is being triggered by release of stored CO2 into the atmosphere by the forests whenever they are degraded or cleared. Thus, carbon accounting within a forest ecosystem and variations resulting from man-made activities, in carbon stocks, is an important phase for significant depiction of forests in climate change strategy at global scale, national scale and local scales (Watson, 2009).

Carbon sequestering is advantageous and important in environmental viewpoint as it includes the important aspect of minimizing CO2 concentration form the atmosphere and making biodiversity richer (Baral, et al., 2009).

Deforestation subsidizes significantly to global CO2 emissions which is considered the second largest cause of GHG discharges in framework of global climate change. To avoid hazardous intrusion in the climate, tropical forest could be pivotal in global determinations to alleviate GHG concentrations. Deforestation and forest degradation results as net causes of carbon emission into the atmosphere mainly due to activities of landuse alteration and forestry on global scale (Mya, 2010).

Because of fossil energy use, deforestation another anthropogenic source, CO2 is accumulating in the atmosphere and is shifting the overall climate (Harries, et al., 2001). Recent understanding of global carbon progression advocates that sequestration of GHGs can be increased through proper forest and agricultural land management (Dixon & Turner, 1991).

Evidence about distribution and expansion AGB is a precondition for appraisal of the role of carbon interchange rates and for convincing prophecies of forthcoming climate deviations (Fuchs, et al., 2009).

Biomass quantification for any forest area is a time-taking activity that is why it is a need to develop convenient, indirect ways for estimating difficult-to-measure variables for biomass quantification (Cordero & Kanninen, 2003).

In the context of international efforts for reducing the atmospheric share of GHGs, United Nations Framework Convention on Climate Change (UNFCCC) and Kyoto Protocol are main two agreements between the countries of the world. The signatory countries of UNFCCC require to cut their human-generated national secretions of CO2 by an average of five percent compared to their emissions in 1990 by 2008-2012 and for this they have to develop a system for estimating carbon stocks for the year 1990 to use as a baseline so that they may reduce their emissions accordingly. In this regard, the United Nations’ (UN) plan of Reducing Emissions from Deforestation and forest Degradation (REDD) is considered an important mitigation plan (Maharajan, 2012).

1.1.1 Biomass estimation and remote sensing

In order to make a plant grow, CO2 is sequestered from the atmosphere. Biomass system is being run by various forms of energy that could be supplied by fossil fuels,
hydroelectricity or might be a part of the biomass system that contributes to the emissions of CO2. The following figure 1.1 presents the concept of biomass generation and CO2 emission.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.1: Biomass system generation and stages of CO2 emission

Developing countries suffer from poor or no forest inventory data to get accurate biomass statistics. With the increasing global sustenance for climate change mitigation efforts and the forthcoming national carbon accounting commitments, the interest for AGB estimation using remote sensing methods has increased rapidly over the recent years (Goetz, et al., 2009). Remote sensing technology is an operational method for managing and observing forest possessions (Kubo & Muramoto, 2008). Techniques in
remote sensing has also upgraded its importance significantly during the last years due to enhanced availability, coverage and finer geometric and spatial resolutions of the remotely sensed imagery (Lu, et al., 2002; Muukkonen & Heiskanen, 2005; Nelson, et al., 1988). These remote sensing techniques also offer the potential to assess biomass at diverse scales. Direct measurement of carbon stock through remote sensing is not possible until present, so additional field inventory data is needed (Rosenqvist, et al., 2003). Dry biomass of species is usually calculated through in-situ measurements as it gives best reliable results but on the other hand this method is quite expensive in terms of cost and also laborious so biomass calculation through remote sensing techniques and methods gives advantage to cover large regions and areas in relatively much efficient in time consumption and also cost effective (Pareta & Pareta, 2011).

Due to compound assemblage, inconstant topography, and great extent of forests, precise estimation of forest carbon stocks is still a trial for both remote sensing and field inventory surveys (Pareta & Pareta, 2011).

1.2 Study Area

Ayubia National Park (ANP) lies in the great Himalayan mountain series which is extended widely to the north of Indian subcontinent. The Himalayan mountain range possesses rich biodiversity and eco-regions for wildlife inhabitants and plant species.

1.2.1 Location, demography and landuse/ landcover

Study area (figure 1.2) with 3312 ha (WWF, 2008) is one of the 21 national parks in Pakistan. ANP lies in Khyber Pakhtunkhwa (KPK) province, former North-West Frontier Province (NWFP), which constitutes 40% of the country’s forested area (Adnan, 2011).

Under category V of IUCN, The ANP falls in one of the fourteen declared National Parks in Pakistan. ANP lies in between 73.40E, 34.11N to 73.41E, 34.01N and is surrounded by the towns of Nathiagali, Ayubia and Khanaspur. ANP was declared as a Protected Area in1984with an area of 1684 ha while later in 1998 its area was extended and re-notified (WWF, 2008).

Climatically, ANP forest lies in Himalayan mountains range in northern areas of Pakistan with moist temperate and moist dry temperate Forest zone. Some of its aspects were present in moist temperate while other in the moist dry temperate (WWF, 2008).

While delineating ANP’s boundary in 2008, WWF - Pakistan observed the forest of ANP as follows;

- The main forest type of ANP was “Conifer forest” with three main species of Fir, Deodar and Blue Pine.
- Conifer forest was mixed with broadleaved forest, different types of shrubs and grasses at low elevation. There were very few patches of pure broadleaved forest in ANP.
- ANP has some large patches of pasture and grasslands.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.2: Study area map of Ayubia National Park (ANP) along with elevation

Abbildung in dieser Leseprobe nicht enthalten

Table 1.1 shows the landuse/ landcover classification of ANP with the help of satellite imagery.

Table 1.1: Landuse/landcover (LULC) classification of ANP (WWFPakistan)

Abbildung in dieser Leseprobe nicht enthalten

In altitude, the ANP ranges from 1,220 m to 2,865 m. The main headquarter of ANP is located in Dunga Gali which is 75 Km from the capital city of Islamabad, Pakistan. An incredible number of tourists come to visit the main town of Nathiagali especially in summers while in winter they come to relish the snowfall. Karalls and Abbasis are the main major ethnic groups of ANP who are distributed in 12 villages located around this park having approximately 50,000 people (Adnan, 2011).

The mean annual rainfall that the parks receive is between 1065 mm - 1424 mm and the mean annual rainfall in between 1 m - 2.5 m (Lodhi, 2007).

1.2.2 Flora

410 species of combined vegetation have been informed from the ANP that belongs to Pteridophytes, Spermatophytes, Lichens and Fungi (Shafique, 2003). The park is mainly comprised of three forest ecotypes including sub-alpine meadow, moist temperate forests and sub-tropical pine forests (Adnan, 2011). The main vegetation consists of dispersed broad-leaved species of tree such as Quercus dilatata (Holly Oak), Ulmus wallichiana, Aesculus indica (Horse chestnut - Ban Khor), and Prunus padus (Kalakat) and mixed with these broad-leaved species there are coniferous tree species of Pinus wallichiana (Blue pine) andAbies pindrow(Fir) exists as well; while other tree species that are rejuvenating in the park are Acerceasium (Maple), Picea smithiana (Spruce), Taxus wallichiana (Yew), and Populus ciliata (Palach) along with Cedrus deodara (known as Deodar) (Lodhi, 2007).

1.2.3 Fauna

Chukors, koklas, flying squirrels, yellow-throated martins, Rhesus monkeys and common leopards are some of the interesting variety wildlife species that ANP possesses (Adnan, 2011). Approximately 16 classes of reptiles, 31 species of mammals, 650 described species of insects, 3 species of amphibians and 23 species of butterflies is reported in ANP (Lodhi, 2007).

1.2.4 Livestock

One buffalo and one goat per household is the average number of animals for the people living around and in the park while the total livestock population in the area is about 27,181 including goats (8,526), cows (4,370), horses (95) and donkeys (114). Warm clothes, carpets and handicrafts are manufactured by using the hair and wool obtained from goats and sheep while for domestic use people have small number of local chickens as well (Adnan, 2011).

1.2.5 Geology and soil

Existing geological structure in the park due to extensive folding, faulting and shearing connected to the local crystal twist ascending from the northward seduction of the Indian sub-continental plate underneath the europium plate having the major rock types of silt, limestone along with metamorphic series of phylites, granite and schist. An average eminence of fir crops and blue pine is supported by reasonable depth of the mineral soil (Adnan, 2011).

Loamy textured shallow soils have exposed bedrock in some grazing areas because of the great incidence of the biotic activities. Rocks in study area belong to Margala Hill limestone which is subordinate and shale (Adnan, 2011).

1.2.6 Tourism impact

Pleasant scenic views, cool weather in summers and snowfall in winters attracts approximately 0.1 - 0.2 million tourists in the towns around and in the park that offer abundant recreational resources. According to an estimate, more than 300 metric tons/ year of fuel wood is being expended by the hotels and vacation houses that provide tourists the facilities for accommodation. While fuel wood and timber requirements are wielding pressure on forest natural resources (Adnan, 2011)

1.3 Research Problem

Forests in the developing country of Pakistan play an important role in the region regarding carbon stock sequestration. Unfortunately, there is no proper system for the carbon stock estimation in the country so it is necessary to develop regional level estimates of carbon stock and fluxes in forests at first. And there is also need of an appropriate methodology for this task to make it possible to assess the potential to preserve and manage forests to grow the accretion of carbon and that project future forest carbon stocks during expected climate and landuse change. Methods in remote sensing have many pluses in AGB assessment compared to traditional ground measurement methods. Measuring carbon stocks through remote sensing is still challenged by the complex canopy structures of the trees in the forests.

1.4 Research Questions

Research questions related to the research problem are:

- How to classify landcover for the study area using an appropriate classification technique?
- How to measure and map carbon stock estimation through combined data of field inventory data and remote sensing data & techniques?
- How to compare carbon stock estimation through linear regression and non­linear & multiple regression model techniques?
-.5 Hypothesis
- Incorporation of remote sensing data and techniques combined with field inventory data is more wide-ranging, reliable, time-efficient and inexpensive as compared to traditional approaches of biomass measurement.
- Non-linear and multiple regression models show best correlations among variables in natural system for carbon stock estimation as compared to linear regression models.

1.6 Research Objectives

Any research without defining objectives is incomplete. Research objectives set directions and dimensions for the researcher to work on. Specified research objectives clarify the purpose of research. Thus, core research objectives for the present study are set as follows:

- Extracting landuse of study area i.e. ANP through high resolution satellite imagery for delineating boundaries of forest area.
- Estimate and regionalize dry AGB and carbon stock for forest area of the study area through combined calculations of field inventories and satellite imagery.
- Applying non-linear and multiple regression models for estimation and regionalization of Above Ground Carbon (AGC). The correlations through linear regression between variables used for carbon stock estimation will be compared to the non-linear multiple regression models.
- Mapping of AGC with an adequate spatial resolution.
- Identifying potential of carbon stock stored in ANP for environmental conservation concerns.
- Identifying role and possible participation of ANP in United Nations project, Reducing Emissions from Deforestation and forest Degradation (REDD).
- Recommendations regarding forest organization of ANP to prevent/ mitigate deforestation and forest degradation.

1.7 Research Significance

Forests in the developing country of Pakistan play an important role in the region regarding carbon stock sequestration. Unfortunately, there is no proper system for the carbon stock estimation in the country so it is necessary to develop regional level estimates of carbon stock and fluxes in forests at first. An attempt to assess forest carbon and biomass fluxes could give lift to attention to authorities towards this issue of organizing rich forests of country.

CHAPTER 2

LITERATURE REVIEW

Chapter 2
LITERATURE REVIEW

The annual altercation of carbon between atmosphere and forests, and the aggregates of carbon stowed in forests, differs commonly with type of forest cover. Five major carbon pools identified and specified by UNFCCC enlisting aboveground, belowground, deadwood, litter and soil. Table 2.1 shows one hundred and one out of 150 developing countries reported that overall carbon stock stored majorly in aboveground (UNFCCC, 2009).

Different parts of tree, originated from physiological or physical interrelations among Table 2.1 Carbon pools identification by 150 countries (UNFCCC)

Table 2.1 Carbon pools identification by 150 countries (UNFCCC)

Abbildung in dieser Leseprobe nicht enthalten

stem sizes, crown extents, foliage area, and biomass amounts, give birth to allometric relationships; and among these tree parts a vital role is being played by foliage in forest growth and is affected by the spatial distribution of the branch biomass (Bartelik, 1996).

The forests are poorly under-inventoried in the developing countries which lead to poor estimates of forthcoming tendencies in these forests (Ozdemir, 2008). A better inventory database is strongly needed which is being produced by low budget inventory methods, comparatively (Gautam, et al., 2004).

The study of (Ragan, et al., 1994) summarizes that response of trees to future climatic conditions will depend on the combined impact of climatic effects to physiological processes and climate-driven shifts in biomass allocation. United Nations program of REDD is trying to combat the climate change challenge. Research done by (Holly, et al., 2007) concludes that the future of REDD along with related climate strategies need not be controlled by the technical encounters of estimating tropical forest carbon stocks as there is a variety of options exists to estimate forest carbon stocks in developing countries and will continue to progress in reaction to the policy needs.

A report published by World Wildlife Fund (WWF) - Pakistan presented the carbon stock assessment in the same area of ANP by (Qasim, 2011) through field inventory. Five carbon pools in ANP were assessed including AGB, Belowground Biomass (BGB), litter, deadwood and soil. A total of 311.69 C t/ha (Carbon tons per hectares) were calculated in ANP combining all carbon pools. Break-up of this carbon stock among carbon pools is presented in table 2.2 as follows:

Table 2.2: Carbon pools estimation in ANP (Qasim, 2011)

Abbildung in dieser Leseprobe nicht enthalten

Through REDD, financial enticements are being provided to assist different developing countries willingly lessen national deforestation tolls and allied carbon discharges under a threshold (based on either a past reference case or prospect projection). The earned credits of carbon storage can then be sold to global carbon marketplace or elsewhere by those countries. These acts and actions can lead us to a better sustainable future combating present and future climate change predictions and biodiversity conservation (Holly, et al., 2007).

2.1 Remote Sensing Data for Carbon Estimation

It is possible to acquire data for sample plotting, site specification and regional mapping through high resolution imagery (<1 m resolution) with the help of aerial
sensors and other new fine resolution sensors of satellite remote sensing. By analyzing such high-resolution imagery, it is possible to get information about forest through extraction of a single distinctive tree information; and there are many approaches used to automatically locate or delineate an individual tree crown, of which most common phenomenon is that trees usually seem as bright objects enclosed by darker shaded areas on high resolution imagery (Donald, et al., 2005). Treetops reflectance shows up extra intensity as compared to the crown periphery while using high-resolution satellite data because of crown outlines and canopy arrangement. The relative intensity of reflectance is represented by the Digital Number (DN) value of the satellite imagery data. So DN values at the periphery of crowns are smaller than at treetops (Hirata, et al., 2009).

Reported by (DeFries, et al., 2006), an approach was suggested in which it was discussed that medium resolution satellite data should be analyzed in order to identify areas or locations that have biomass/ carbon hotspots and then those hotspots should be observed and analyzed with more high-resolution satellite images in order to get more detailed and precise statistics. This approach will save time, cost and effort and will also reduce the need to analyze the data. The diagrammatic representation of this approach has been presented in figure 2.1.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2.1: A conceptual approach for monitoring biomass in large scale (DeFries, et al., 2006)

Study by (Rahman, et al., 2005) describe that remote sensing tools and techniques are useful in estimating forest biomass based on correlation in between different variables extracted from satellite data of Landsat having Enhanced Thematic Mapper (ETM+) sensor. Their study used a number of dummy variables to increase the correlation between the predictor variables for estimating biomass in the forest sited in southern Chittagong of Bangladesh. They used stratified random sampling to layout their plots for field samples for diverse vegetation types in study area. They used diameter at breast height (DBH) along with height of the tree taken as key indicators to estimate the biomass in the sample plots. They converted these key indicators for into biomass based on allometric relationships that were available for some species and for other species they converted these key indicators into biomass using conversion constants. Simple and multiple regressions were used in this study to predict the biomass in the study area between different spectral satellite bands of Landsat ETM+ sensor. Low correlations were found between variables as the value of correlation (r2) was found in band 2 and band 4 of Landsat ETM+ that was r2<0.26 and similar correlation was observed in multiple regression as well. Even satellite data spectral transformation (vegetation indices, principal components etc.) could not develop a strong correlation. When introduced dummy variables the correlation increased radically and the value of r2 was obtained r2=0.939. The study concluded that visible band 2 of satellite sensor data of Landsat ETM+ contains maximum information on forest biomass as observed through regression analysis incorporating the dummy variables.

(Japhet, et al., 2013) presented a research study on forest cover change, quantification and mapping of above ground carbon stock using the techniques and tools in GIS and Remote Sensing with incorporation of forest inventory data. They used Landsat images for their study area of year 1980, 1995 and 2010 taken in dry season to estimate the landcover changes. Using remotely sensed data of satellite imagery in dry season was also advantageous in respect that it minimizes the seasonality and cloud cover in the images. They classified the Landsat images in Erdas Imagine software for remote sensing image interpretation and image analysis using the Maximum Likelihood Classifier and then calculated area of their each landcover class in ArcGIS software and exported final excel sheet from ArcGIS software to make the graphical charts of their results. The landcovers of vegetation obtained after image classification were closed forest, open forest, bushland and grassland and other landuse classes (e.g. bare land, settlements, cultivated land). For the sake of field inventory plots in which they used concentric plots and sub-divided each plot in to sub-plots (concentric) of radius 2m, 5m, 10m and 15m. The two main parameters for carbon stock assessment were recorded including tree’s DBH and tree height using veneer caliper and Suunto hypsometer respectively. The trees with DBH <2cm and >1cm were recorded in 2m radius of concentric plot, trees with DBH >=2cm and <10cm were recorded in 5m radius, trees with DBH >=10cm and <20cm were recorded in 10m radius while trees with DBH >=20cm were recorded in 15m radius of the concentric plots. A product of total tree volume and wood basic density was used to compute the tree AGB. Further conversion of biomass into carbon was done using the conversion factor of 0.49. The carbon mapping was done through Ordinary Kriging using exponential semi variogram model in the ArcGIS software as this method has the finest performance for AGB estimation and observing its spatial heterogeneity. Their study results showed that overall classification accuracy of their study area was above 80%. The results also revealed that closed forest area has maximum carbon stock density as compared to other vegetation classes in the study area. This study also examined an extensive change in forest cover between the period 1980 and 2010. Potential estimated carbon storage of 7206.46 tC and 14730.41 tC was obtained in two regions of study area.

In a review article (Dengsheng, 2006) has presented a review of satellite based, GIS based and field based methods for estimating AGB/ carbon stock. In the article the author has discussed that through SPOT satellite data we can estimate carbon stock on per-pixel. Table 3 present the characteristics of other data sets used for carbon stock approximation.

Biomass monitoring in forest ecosystems is important as forests are diminishing rapidly in many parts of the world, which is one of the chief sources of global carbon emission. Remote sensing is a useful technique for rapid estimation of biomass. Many studies are currently available to assess biomass from satellite sensor data using regression provide low correlation. The study reconnoiters the possibilities to increase it. Numerous spectral channels and alterations of Landsat Enhanced Thematic Mapper Plus (ETM+) data for forecasting biomass in a tropical forest ecosystem of south­eastern Bangladesh were tested. One of the interesting verdicts of the study is the incorporation of dummy variables based on forest types can radically increase the correlation. Numerous techniques and methods for estimating aboveground biomass has been summarized in following table 2.3.

[...]

Excerpt out of 79 pages

Details

Title
Estimating Forest Tree Carbon using Remote Sensing Data and Techniques
College
University of the Punjab  (Institute of Geology)
Course
Geomatics
Grade
3.87
Author
Year
2013
Pages
79
Catalog Number
V372218
ISBN (eBook)
9783668500259
ISBN (Book)
9783668500266
File size
12637 KB
Language
English
Tags
Carbon Emission, Ayubia National Park, Remote Sensing, Geographical Information System (GIS), SPOT-5, Landsat, Aboveground Carbon (AGC), Aboveground Biomass (AGB)
Quote paper
Adeel Ahmad (Author), 2013, Estimating Forest Tree Carbon using Remote Sensing Data and Techniques, Munich, GRIN Verlag, https://www.grin.com/document/372218

Comments

  • No comments yet.
Read the ebook
Title: Estimating Forest Tree Carbon using Remote Sensing Data and Techniques



Upload papers

Your term paper / thesis:

- Publication as eBook and book
- High royalties for the sales
- Completely free - with ISBN
- It only takes five minutes
- Every paper finds readers

Publish now - it's free