Ahmad Fikri Abdullah

Agriculture land suitability analysis evaluation based multi criteria and GIS approach

Land suitability evaluation (LSE) is a valuable tool for land use planning in major countries of the world as well as in Malaysia. However, previous LSE studies have been conducted with the use of biophysical and ecological datasets for the design of equally important socio-economic variables. Therefore, this research has been conducted at the sub national level to estimate suitable agricultural land for rubber crops in Seremban, Malaysia by application of physical variables in combination with widely employed biophysical and ecological variables. The objective of this study has been to provide an up-to date GIS-based agricultural land suitability evaluation (ALSE) for determining suitable agricultural land for Rubber crops in Malaysia. Biophysical and ecological factors were assumed to influence agricultural land use were assembled and the weights of their respective contributions to land suitability for agricultural uses were assessed using an analytic hierarchical process. The result of this study found Senawang, Mambau, Sandakan and Rantau as the most suitable areas for cultivating Rubber; whereas, Nilai and Labu are moderately suitable for growing rubber. Lenggeng, Mantin and Pantai are not suitable for growing rubber as the study foresaw potential environmental degradation of these locations from agricultural intensification. While this study could be useful in assessing the potential agricultural yields and potential environmental degradation in the study area, it could also help to estimate the potential conversion of agricultural land to non-agricultural uses.

A Methodology for Processing Raw LiDAR Data to Support Urban Flood Modelling Framework: Case Study—Kuala Lumpur Malaysia

High quality representation of the topographic and the correct representation of significant urban features would be a fundamental foundation to a better urban flood model. Without such a representation, simulation of flood behaviours would be less successful as the flow patterns were completely dependent on ground levels and the shape of the features. Typically, such data can be obtained via Light Detection and Ranging (LiDAR) surveys. The process of turning raw LiDAR data into a useful Digital Terrain Model (DTM) involves careful processing and application of thinning, filtering and interpolation algorithms. Filtering is a process of automatic detection and interpretation of bare earth and objects from the point cloud of LiDAR data, which results in the generation of a DTM. To date, many filtering algorithms have been developed, and in a more general sense, many of them have become standard industry practice. However, when it comes to the use of a DTM for urban flood modelling applications, these algorithms cannot be always considered suitable. Depending on the terrain characteristics, they can even lead to misleading results and degrade the predictive capability of the modelling technique. This is largely due to the fact that urban environments often contain a variety of features (or objects) such as buildings, elevated roads, bridges, curbs and others which have the ability to store or divert flows during flood events. As these objects dominate urban surfaces, appropriate filtering methods need to be applied in order to identify such objects and to represent them correctly within a DTM so that the DTM can be used more safely in modelling applications. The work described in this chapter concerns improvements of a LiDAR filtering algorithm. The key characteristics of this improved algorithm are: ability to recover curbs and the use of appropriated roughness coefficient of Manning’s value to represent close-to-earth vegetation (e.g. grass and small bush). The results of the improved algorithm were demonstrated using Kuala Lumpur (Malaysia) as a case study. Improvement, in terms of a difference in flood depths and flood flows were observed between the hydraulics models built from several available filtering algorithms and the improved algorithm (MPMA). The overall results suggest that the improvement made in MPMA can lead to some difference in model results, which may in some cases be significant with a tendency towards incorrect flood flow by those models in which such features are not properly represented.

Quantification of terrestrial laser scanner (TLS) elevation accuracy in oil palm plantation for IFSAR improvement

In order to ensure the oil palm productivity is high, plantation site should be chosen wisely. Slope is one of the essential factors that need to be taken into consideration when doing a site selection. High quality of plantation area map with elevation information is needed for decision-making especially when dealing with hilly and steep area. Therefore, accurate digital elevation models (DEMs) are required. This research aims to increase the accuracy of Interferometric Synthetic Aperture Radar (IFSAR) by integrating Terrestrial Laser Scanner (TLS) to generate DEMs. However, the focus of this paper is to evaluate the z-value accuracy of TLS data and Real-Time Kinematic GPS (RTK-GPS) as a reference. Besides, this paper studied the importance of filtering process in developing an accurate DEMs. From this study, it has been concluded that the differences of z-values between TLS and IFSAR were small if the points were located on route and when TLS data has been filtered. This paper also concludes that laser scanner (TLS) should be set up on the route to reduce elevation error.

A hybrid model using machine learning methods and GIS for potential rockfall source identification from airborne laser scanning data

The main objectives of this paper are to design and evaluate a hybrid approach based on Gaussian mixture model (GMM) and random forest (RF) for detecting rockfall source areas using airborne laser scanning data. The former model was used to calculate automatically slope angle thresholds for different type of landslides such as shallow, translational, rotational, rotational-translational, complex, debris flow, and rockfalls. After calculating the slope angle thresholds, a homogenous morphometric land use area (HMLA) was constructed to improve the performance of the model computations and reduce the sensitivity of the model to the variations in different conditioning factors. After that, the support vector machine (SVM) was applied in addition to backward elimination (BE) to select and rank the conditioning factors considering the type of landslides. Then, different machine learning methods [artificial neural network (ANN), logistic regression (LR), and random forest (RF) were trained with the selected best factors and previously prepared inventory datasets. The best fit method (RF) was then used to generate the probability maps and then the source areas were detected by combining the slope raster (reclassified according to the thresholds found by the GMM model) and the probability maps. The accuracy assessment shows that the proposed hybrid model could detect the potential rockfalls with an accuracy of 0.92 based on training data and 0.96 on validation data. Overall, the proposed model is an efficient model for identifying rockfall source areas in the presence of other types of landslides with an accepted generalization performance.

Improvement of Digital Elevation Model (DEM) using data fusion technique for oil palm replanting phase

ABSTRACT Digital elevation models (DEMs) play an important role in producing terrain-related applications such as curvature and contour maps for planning and management of oil palm plantation. Data fusion of DEMs derived from terrestrial laser scanning (TLS) and interferometric aperture radar (IfSAR) was developed with the intention to increase the accuracy of IfSAR-derived DEM at a lower cost thus, provide a high-quality data for plantation management. In this research, fusion by weights was carried out after applying regression analysis to integrate both TLS and IfSAR data. The results showed a significant reduction in root mean square error (RMSEs) after fusion. RMSEs of both stations reduced from 1.83 m to 0.35 m and from 3.13 m to 0.41 m for Station 1 and Station 2, respectively. In addition, data fusion technique for an area with no TLS data were tested around the stations at 200 m distance. The RMSEs decreased from 2.52 m to 2.33 m for Station 1 but the value increased from 2.09 m to 2.13 m for Station 2. It was concluded that the proposed fusion technique in the extension area could be done in a relatively flat area but not be used in a steep-slope area.

Estimation of soil loss in Seremban, Malaysia using GIS and remote sensing technique

Runoff causes soil loss and is a continuous ecological problem in Seremban, Malaysia. It is crucial to collect data on soil loss for improved agricultural productivity and to manage natural resources effectively. This research maps the distribution and estimates the yearly mean value of soil erosion through the utilization of techniques of remote sensing and GIS by implementing the Revised Universal Soil Equation (RUSLE). To determine the variables of RUSLEs soil loss and analyze them in an integrated GIS environment, we used a scale of 1:50,000 according to criteria of topographic map, Aster Digital Elevation Model (DEM) which has a feature of spatial resolution that extends up to 20 m, a soil map which is digitally programmed with a scale of 1:250,000, and a decade of rainfall records for 12 stations. The data revealed that Seremban records an annual soil loss that ranges from no soil loss in forested areas (Lenggeng - Panti - Ampangan - Seremban) to >100 tone hectare per year in the open area ((Labu - Renggam - Lenggeng). The total annual soil loss is estimated at 883 tonnes/hectare/year and is distributed across different land cover as follows: 198 tonnes from agriculture areas, 39 tonnes from forest areas, and 20.45 from rural areas, 610 tonnes from open area, 12 tonnes from urban areas, and 1.4 tonnes from inland water areas.

Monitoring spatial and temporal variations of the rice backscatter coefficient (σ0) at different phenological stages in Sungai Burong and Sawah Sempadan, Kuala Selangor.

Monitoring rice growth and yield estimation using optical remote sensing data constitutes a big challenge largely due to cloud conditions that are typical of tropical regions. Using Radar remote sensing data helps because it overcomes the cloud issue and distinguishes the behaviour of the radar backscattering of rice crops specifically. This study indicated the temporal change of rice backscatter (σ°) at two different growth stages using HH polarimetric Radarsat-2. The aims of this study are: (1) to identify crop with different life spans based on the backscatter coefficients values from a single polarisation for understanding the backscatter characteristic of rice over the entire growth cycle, and (2) to understand the advantages and limitations using the RADARSAT-2, C band with HH polarisation. The values of backscattering coefficients have been related to the Malaysia rice crop calendar to get the information of the growth status. The result shows strong backscatter coefficient values on the 21st of May that referred to the reproductive-maturity of rice in the Sawah Sempadan area, and out of season for the Sungai Burong area. While for the August 1st imagery, the result shows weak backscatter values which refers to early vegetative and vegetative-reproductive. The values of backscattering coefficient are found to be much less for early vegetation compare to mature rice crop. In this paper, we have also performed a classification of a rice field using Landsat 8 OLI.