This dissertation is submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy. I would like to thank all members of the UAEU College of Science, especially the Department of Biology.
Introduction
Overview
Furthermore, none of these equations have been developed and applied to fit one of the most important dryland fruit crops, the date palm Phoenix dactylifera. Indeed, the date palm appears to be the only native wild desert plant domesticated in its native harsh environments (Zohary & Hopf, 2000).
Statement of the Problem
Aim and Objectives
Literature Review
- Phoenix dactylifera, Date Palm
- Quantifying Terrestrial Carbon Sequestration
- Biomass Allometric Equations
- Geospatial Technologies for Estimation of Carbon Stock
- Arid Lands Case Studies
- Learning Lessons from the Literature Review
The mathematical model commonly used for modeling AGB is based on the power function (Yuen et al., 2016). The contrast between these channels can be used as an indicator of the greenness of the vegetation (Xie et al., 2008, p. 200).
Structure of the Dissertation
The chapter concludes with a date carbon stock estimation map of Abu Dhabi.
Research Methodology
Study Area
Geographic Settings
In the summer season, there is a high incidence of suspended matter throughout the country brought by the prevailing wind from the head of the Arabian Gulf (Western, 1989). Based on soil properties of the desert soil, "Salting" is inferred as the dominant mechanism of soil particle movement, followed by surface creep and the suspension movement (Shahid & Abdelfattah, 2008). In general, biological activity in local soils is very low, and only about three percent of the entire land is naturally suitable for arable cultivation (Al-Rawai, . 2004).
Remote Sensed Data
The vector boundary shape file of the Emirate of Abu Dhabi was used to subdivide the study area. Mosaic operation based on weighted seam line generation procedure was applied for scenes and 161/44 as these images involve the main urban centers of the study area, while for scenes: 162/43 and 162/44 the geometry based seam line generation procedure was applied as the images contain homogeneous areas such as desert and sabkhas (Figure 11), (Al Ahbabi, 2013). The pan-sharpening of WV-2 multispectral bands was achieved using the NN algorithm to produce images with 0.50 m pixel size (Jawak et al., 2013).
Methodology Flowchart
Developing Allometric Equations for Date Palm
- Field Data Collection
- Field Measurements
- Determination of Allometric Equations
Although some researchers prefer to use the fresh weight instead of dry weight to build their equations (Dewi et al., 2009; Khalid et al., 1999a) (Appendix 1). Samples were weighed to obtain the percentage of dry weight to original fresh weight in each sample (dry to fresh factor = DF) (Figure 14). Trunk Dry Weight = Trunk Fresh Weight × Trunk DF Root Dry Weight = Root Fresh Weight × Root DF Percentage of BGB.
LULC Classification and Accurate Mapping of
- LULC Classification
- Mapping Vegetated Area & Delineating of Date Palm
- Mapping Young, Medium, and Mature Date Palm Plantations
- Assessing Accuracy
Finally, the thematic LULC map was created and the area of each of the seven classes was calculated in hectares (Chapter 5, subsection 5.2.1). The shapefile will then be used for the selection of the corresponding WV-2 scenes covering the vegetated areas present in the study area. A confusion matrix was produced and accuracy metrics were calculated for each class of the LULC map, as well as DP on different age-stage maps (Chapter 5, subsection 5.2.4).
Building Remote Sensing based Models for Biomass and
- Field Data Collection
- Identifying Remote Sensing Variables (Predictors)
- Statistical Analysis
- Models Evaluation
- Applying the Remote Sensing Based Models to Estimate
- Calculating the Table Carbon Stock of Date Palm Plantations
Scatter plots were drawn to visualize the relationships between field-estimated AGB of DP correlated with RS predictors (see Chapter 4, Subsection 4.2.3 and Chapter 6, Subsection 6.2.3). Maps of DP of Abu Dhabi generated from a previous study using submeter WV-2 imagery were used (see Chapter 2, Subsection 2.6.3 and Chapter 5, Subsection 5.2.3). Therefore, the SOC in tons of DP for the three age stages was calculated according to equation 10.
Development of Date Palm Biomass Allometric Equations and
Overview
In addition, such methods are tedious and time-consuming (Ebuy et al., 2011) and therefore cannot be used routinely. Subsequently, allometric equations have been developed and used to estimate tree biomass and CS from dendrometric measures such as tree diameters and height (Ebuy et al., 2011; . Picard et al., 2012). More recently, allometric equations have been used, coupled with RS and field-based structural variable measurements (Fonton et al., 2017; Salem Issa et al., 2019).
Results
- Field Variables of Date Palm at Different Age-Stages
- Ratios of Date Palm Biomass Components
- Biomass Allometric Equations of Date Palm
- Carbon Stock in Date Palm Plantations at Different
The contribution of the crown and root to the total biomass increased with age, thus with trunk growth of the palm. The ratio of CB to total biomass decreased to 35.75% and 34.89% of the total biomass for medium and mature DP, respectively. This percentage represents only about two-thirds of that recorded from samples taken under ("In") the date palms (Table 15).
Summary
Furthermore, the average ratios of belowground biomass (BGB) to aboveground biomass (AGB) varied with palm maturity stages on average and 0.496 for young, middle and mature palms, respectively. Furthermore, the results showed that the amounts of organic carbon (OC) stored in dates were significant with values of: 15.88 kg. Significantly higher amounts of SOC were measured compared to other local plants with values: 18,092 kg.
A Pilot Study to Assess Carbon Stock in Date Palm Plantations
Overview
RS data is correlated with plot-based field measurements to estimate AGB, and therefore CS. 3 of the thesis by characterizing the carbon stock of date palm using an RS-based biomass model (see Chapter 1, Subsection 1.3 Purpose and Objective). This chapter specifically focuses on: (1) Identifying the most reliable RS variables to estimate the AGB of DP in Abu Dhabi using Landsat 8 OLI images, (2) Building an RS-based biomass model to CS in DP in the study area.
Results
- Date Palm Plantations’ Structure and Plot Densities
- The Field-Based Biomass Estimation Model
- The RS-Based Biomass Estimation Model
For middle and young DP, none of the Landsat RS variables showed significant correlations with AGB (except with one band, SWIR1 and the TCB index only for middle DP) (Appendix 5). The stepwise regression results showed that the linear regression model combining individual Landsat 8 OLI (G, NIR and SWIR2) and VI (DVI, NDGI and RVI) bands improved the R² value for AGB prediction for DP and thus gave more accurate prediction results biomass and CS (equations 19 and 20). CS map for the best model of the study area produced using the spatial modeling tool in the ERDAS software.
Summary
However, for the medium and young DP (10–5 and less than 5 years) the correlation was not significant (with the exception of SWIR1 and TCB index for medium DP), where the use of higher spatial resolution should be a good alternative to be (see Chapter 5); in addition to expanding the actual field plots to include more plots representing all three age stages of DP (see Chapter 6). A stepwise multiple regression analysis combining DVI, NDGI and RVI vegetation indices improved the value of the R2 to a value of 0.952. In order to accurately visualize the results on maps of DP's KS, an accurate mapping of DP plantations (at different age stages) is obtained in the subsequent Chapter 5.
LULC Classification of Abu Dhabi and Accurate Mapping of
Overview
Furthermore, for any algorithm to work properly, crowns must, at a minimum, be detectable and segmented as an object in the image prior to classification. Using the pan-enhanced WV-2 images (spatial resolution 0.5 meters) (see Chapter 2, Subsection 2.3), DP crowns can be distinguished from the background (soil, grasses and weeds) and other shrubs and trees by color , tones to be used , texture, size and plant arrangement (Figure 27). The DP can be visually distinguished from other vegetation (grasses, trees and shrubs) using the tools mentioned plus the plant arrangements and spacing.
Results
- LULC Map Creation Using Hybrid Classification
- Mapping Vegetation and Date Palm Using Landsat-8 OLI
- Mapping Young, Medium, and Mature Date Palm
- Maps Validation
Most of the DP plantations are shown to be found in AlAin (right box) and Liwa (left box). It can be noted that more than half of the Abu Dhabi DP plantation areas were mature DP (> 10 years). The accuracy of the DP map derived from WV-2 was also assessed for all three age stages combined (considered only as one DP class).
Summary
The total number of DP planted in the study area numbered an estimated number of palms (Table 24).
Remote Sensing Based Models for Assessing Date Palm
Overview
Specifically, this chapter presents (1) the construction of an RS-based spatial model for estimating the biomass and CS of DP, and (2) the quantification and visualization of the amount of biomass and CS in Abu Dhabi, using the spatial models built based on RS.
Results
- Descriptive Statistics of the Field Variables Assessed
- The Aboveground Biomass Analysis and
- Models Development and RS Variables Importance
- Modeld Validation
- Creation of the AGB Map of DP and Estimating Total Carbon
Furthermore, the results of stepwise multiple regression analysis on AGB of mature PD showed that a combination of single bands or VIs does not improve R2. Therefore, the second-order polynomial equation using only TCW as the RS predictor is the strongest model to estimate mature DP biomass with R² equal to 0.7643 and P value equal to 0.007 (Equation 21 and Figure 35) . The results of stepwise regression analysis on AGB of immature PD showed that a combination of single bands or a combination of VIs does not improve R2.
Summary
For medium DP, the second-order polynomial equation using only DVI as RS predictor is the strongest model to estimate the biomass of medium DP with R² equal to 0.2286 and P value equal to 0.049. While for young DP, the linear regression equation using only NIR as RS predictor is the strongest model to estimate the biomass of young DP with R² equal to 0.2828 and P value equal to 0.023. An exponential regression equation using RDVI as RS predictor was the best single VI and had the strongest correlation among all RS variables from Landsat 8 OLI for AGB of non-mature DP, with an R2 value of 0.4987 and P value equal to to 0.00002.
Discussion
Due to the reasons explained in Chapter 1 (Subsections 1.4.4.3), the moderate resolution Landsat-8 OLI imagery was chosen. The results of the regression analysis on the AGB of the middle DP and the new DP in the pilot study (see Chapter 4) revealed that none of the Landsat RS variables showed any significant correlation with the AGB (except with the band of only, SWIR1, and TCB Index only for medium DP). Most of the PD plantations in Abu Dhabi were found in Al Ain (in the east of the emirate) and Liwa (in the south of the emirate) with more than half of them in the mature stage (> 10 years).
The dominant role of semi-arid ecosystems in the trend and variability of the land CO2 sink. Review of the use of remote sensing for biomass estimation to support renewable energy generation.
Conclusion DQG Recommendations
