chek jawa map
introduction
methods & data
analysis
discussion
Grain size distribution, Chek Jawa
February 2008
distance from NE shoreline
distance from SE shoreline
distance from coastline
% seagrass coverage
in sandbar
region
flood tide currents
ebb tide currents
1000 ≤ w < 2000
500 ≤ w < 1000
355 ≤ w < 500
250 ≤ w < 355
125 ≤ w < 250
100 ≤ w < 125
63 ≤ w < 100
45 ≤ w < 63
w < 45
Introduction
To see the world in a grain of sand,
and to see heaven in a wild flower, hold infinity in the palm of your hands,
and eternity in an hour. - William Blake
and to see heaven in a wild flower, hold infinity in the palm of your hands,
and eternity in an hour. - William Blake
Background
Study area
Beach sediments and
Grain size sorting
Grain size sorting
Objectives
The understanding of nearshore processes is fundamental to the preservation and protection efforts of beaches and coastal assets (Komar, 1998). The processes governing the movement of beach sediment has always been an area that has intrigued both academics and governing bodies due to their importance in affecting the variability in beach morphology. In this instance, the focus will be on the grain size sorting process on the beach.
Following studies done by Evans (1939), Komar (1977) and Self (1977) on the grain size sorting and selection process on beaches, this website documents the application of Geographical Information Systems (GIS) in the study of the grain size sorting process of beaches, taking as a case example the protected beach of Chek Jawa, which is on the island of Pulau Ubin, Singapore.
Following studies done by Evans (1939), Komar (1977) and Self (1977) on the grain size sorting and selection process on beaches, this website documents the application of Geographical Information Systems (GIS) in the study of the grain size sorting process of beaches, taking as a case example the protected beach of Chek Jawa, which is on the island of Pulau Ubin, Singapore.
The beach of Chek Jawa lies on the eastern tip of Pulau Ubin, and has a unique position at the confluence of the Johor River and the Straits of Johor (to see, click "regional view" in map view).
A small coast of roughly one square kilometer in area, Chek Jawa has a total of six major habitats, namely coastal forest, sandy beach, mangrove, lagoon, rubble-strewn coral flat and a tiny island. Chek Jawa is currently protected and managed by the National Parks Board of Singapore after the decision to defer from land reclamation in 2001.
A small coast of roughly one square kilometer in area, Chek Jawa has a total of six major habitats, namely coastal forest, sandy beach, mangrove, lagoon, rubble-strewn coral flat and a tiny island. Chek Jawa is currently protected and managed by the National Parks Board of Singapore after the decision to defer from land reclamation in 2001.
Beach sediments typically consist of sand or gravel particles of various sizes, the proportion of which can be determined by granulometric (grain size) analysis and then further classified into grain size categories (Bird, 2000) via sieving methods.
The distribution of the grain sizes on a beach would reflect the underlying surface topography of the beach as well as the local intensity of wave energy and its dissipation (Komar, 1998). These would provide an insight on the nature of the beach environment and in the case of Chek Jawa, aid in the study and understanding of this unique coastal habitat.
The knowledge of the cross-shore and long shore variations in the proportions of grain sizes can also be compared with other variables to examine their influences on the grain size sorting process.
The distribution of the grain sizes on a beach would reflect the underlying surface topography of the beach as well as the local intensity of wave energy and its dissipation (Komar, 1998). These would provide an insight on the nature of the beach environment and in the case of Chek Jawa, aid in the study and understanding of this unique coastal habitat.
The knowledge of the cross-shore and long shore variations in the proportions of grain sizes can also be compared with other variables to examine their influences on the grain size sorting process.
This project seeks to gain an insight into the coastal processes affecting the beach at Chek Jawa, in particular the grain size sorting process and the variables that influence this phenomenon. This will be achieved by the use of GIS mapping capabilities and statistical analysis to show the phenomenon of selective cross-shore and long shore transportation of different grain sizes. The data yielded would aid in a better understanding of the area and so help in making informed management decisions on this fragile ecosystem.
Furthermore, the applicability of GIS in the study of nearshore processes, in this case the grain size sorting process, of the beach has to be shown to be effective. This would be examined in the areas of data collection and analysis. The primary objective is thus to show that GIS aids in data collection, visualization and manipulation. Also, that the use of GIS gives the potential for multivariate extensions to such projects.
Last but not least, we hope to engage the public in having a general appreciation of natural processes, using an accessible, non-proprietary, interactive GIS built on Google Maps.
Furthermore, the applicability of GIS in the study of nearshore processes, in this case the grain size sorting process, of the beach has to be shown to be effective. This would be examined in the areas of data collection and analysis. The primary objective is thus to show that GIS aids in data collection, visualization and manipulation. Also, that the use of GIS gives the potential for multivariate extensions to such projects.
Last but not least, we hope to engage the public in having a general appreciation of natural processes, using an accessible, non-proprietary, interactive GIS built on Google Maps.
Methods and Data
Even Castles made of sand, fall into the sea, eventually. - Jimi Hendrix
Sampling method
Sieving method
Interpolation method
As Chek Jawa is approximately 1 km by 660 metres, sampling plan and sampling frame has to be thought out carefully so that the interpolation of a logistically possible amount of samples would be as close to the actual scenario as possible.
The distribution of the grain sizes along the shore is affected by many physical variables. Wind direction, the longshore drift and strength of incoming waves affect the distribution of grain sizes on the coast. However, two general trends had been observed about the distribution of grain sizes along the coast. Firstly, largest sediment particles are generally located in the zone of most intense wave breaking and decreases in grain size toward deeper waters and shoreward across the surf and swash zones. (Komar, 1977) Secondly, finer grain sizes tend to be further downcoast in the direction of the longshore drift. This implies that the grain sizes are not equally distributed along the coast. Hence the probability of obtaining a sample of a particular grain size will vary across the coast.
Simple random sampling requires that each possible sample that is to be obtained from a population has to have the same probability of being selected (Petersen and Calvin, 1965). However, based on the two general trends mentioned above, it can be inferred that the probability of selecting any sample with a particular grain size will not hold true. There are many possible variations which may not be detected due to possible clustering of sampling points. Therefore we avoided simple random sampling.
Systematic sampling is another popular method used in the field of soil sampling. This method requires the field to be divided equally such that each unit in the field is regularly spaced out from each other (Cochran, 1953). It requires the sampler to obtain samples from all the strata at a close-enough distance to capture variations, which implies investment of a certain amount of time and effort. Judging from low tide time window for us to collect our samples, we decided that based on these limitations systematic sampling is not feasible due to time and logistical constraints. Another possible disadvantage is a probability of a zero estimate if the sampling interval between systematic samples is larger than the possible period of occurence of significant variations (Christman, 2000; cited in Loh, 2008).
Within these constraints, we used stratified random sampling for our data collection (white squares in "toogle analysis"), based on transect lines which were used in earlier studies, so that it is possible for our raster layer to be superimposed with past data with minimal error from interpolation. Three random sampling points were generated for each 100 metres. This method is used considering time and logistical limitations for data collection and sieving; importantly, it is a compromise between random and systematic sampling, to mitigate extreme clustering and systematic bias.
The distribution of the grain sizes along the shore is affected by many physical variables. Wind direction, the longshore drift and strength of incoming waves affect the distribution of grain sizes on the coast. However, two general trends had been observed about the distribution of grain sizes along the coast. Firstly, largest sediment particles are generally located in the zone of most intense wave breaking and decreases in grain size toward deeper waters and shoreward across the surf and swash zones. (Komar, 1977) Secondly, finer grain sizes tend to be further downcoast in the direction of the longshore drift. This implies that the grain sizes are not equally distributed along the coast. Hence the probability of obtaining a sample of a particular grain size will vary across the coast.
Simple random sampling requires that each possible sample that is to be obtained from a population has to have the same probability of being selected (Petersen and Calvin, 1965). However, based on the two general trends mentioned above, it can be inferred that the probability of selecting any sample with a particular grain size will not hold true. There are many possible variations which may not be detected due to possible clustering of sampling points. Therefore we avoided simple random sampling.
Systematic sampling is another popular method used in the field of soil sampling. This method requires the field to be divided equally such that each unit in the field is regularly spaced out from each other (Cochran, 1953). It requires the sampler to obtain samples from all the strata at a close-enough distance to capture variations, which implies investment of a certain amount of time and effort. Judging from low tide time window for us to collect our samples, we decided that based on these limitations systematic sampling is not feasible due to time and logistical constraints. Another possible disadvantage is a probability of a zero estimate if the sampling interval between systematic samples is larger than the possible period of occurence of significant variations (Christman, 2000; cited in Loh, 2008).
Within these constraints, we used stratified random sampling for our data collection (white squares in "toogle analysis"), based on transect lines which were used in earlier studies, so that it is possible for our raster layer to be superimposed with past data with minimal error from interpolation. Three random sampling points were generated for each 100 metres. This method is used considering time and logistical limitations for data collection and sieving; importantly, it is a compromise between random and systematic sampling, to mitigate extreme clustering and systematic bias.
The properties of the soil are very much determined by the composition of the sizes of its particles. By separating a soil sample, we can obtain its composition. Today many methods exists to segregate the soil particles, to name a few, sieving, sedimentation and pipette sampling methods. The sieving method is amongst the simplest and most commonly used.
Sieving method involve vigorously shaking of sets of differently sized sieves to separate the sample into various grain sizes. It is a convenient method for segregating particles coarser than 0.05mm. (Day,1965) We adopted the recognized Udden-Wentworth grain size classification, which takes into account the characteristics of terrigenous sediments (Wentworth, 1922) (see figure below).
Based on the Udden-Wentworth grain-size classification, we expected that most of our sample grain size to fall within 500 μm to 65 μm, and selected sieves with the following gap sizes: 2mm, 1mm, 500m, 320 μm, 250 μm, 125 μm, 100 μm, 65 μm, 45 μm and sieving pan (to collect grain sizes less than 45 μm ) for segregating the various grain sizes of our samples.
For particle-size analyses, the distribution of a soil is expressed in portions of the various sizes of particles which it contains. The proportions are commonly represented by relative weights of such classes (Day, 1965).
Samples were dried at 400 degrees Celcius for approximately 3hrs to remove excessive moisture content that is held within the sample. This will ensure that all the samples will have approximately the same amount of moisture when it is sieved. A mechanical sieve shaker to ensure that each sample will be subjected to the same length of time and amplitude of shaking. Sieve brushes were used to clean the sieves before segregating the next sample. This is to ensure that earlier segregated samples do not contaminate the later samples as well as keep the sieves dry so as not to collate the sediments and interfere with the results of the segregation.
The sieving method has several limitations. The probability of a particle passing a given sieve in a given time of shaking depends upon the nature of the particle and properties of the sieve (Day, 1965). The dimensions of the particle might hinder the particle from passing through a given sieve opening, even though the particle has the correct mean size to pass though. Thus it will require extensive shaking before the particle is orientated into the “correct” orientation to pass through the sieve opening. Sieve openings are generally unequal in size, requiring extensive shaking before all particles have the opportunity of approaching the largest openings. A "complete" sieving can rarely be met in practical times of shaking, thus we decided to designate our own shaking time (15 minutes). We obtained the 15 minutes sieving time by performing a few trials sieving to see what is the minimum time required to achieve sufficient segregation of the samples without compromising the quality of the data.
Sieving method involve vigorously shaking of sets of differently sized sieves to separate the sample into various grain sizes. It is a convenient method for segregating particles coarser than 0.05mm. (Day,1965) We adopted the recognized Udden-Wentworth grain size classification, which takes into account the characteristics of terrigenous sediments (Wentworth, 1922) (see figure below).
Based on the Udden-Wentworth grain-size classification, we expected that most of our sample grain size to fall within 500 μm to 65 μm, and selected sieves with the following gap sizes: 2mm, 1mm, 500m, 320 μm, 250 μm, 125 μm, 100 μm, 65 μm, 45 μm and sieving pan (to collect grain sizes less than 45 μm ) for segregating the various grain sizes of our samples.
For particle-size analyses, the distribution of a soil is expressed in portions of the various sizes of particles which it contains. The proportions are commonly represented by relative weights of such classes (Day, 1965).
Samples were dried at 400 degrees Celcius for approximately 3hrs to remove excessive moisture content that is held within the sample. This will ensure that all the samples will have approximately the same amount of moisture when it is sieved. A mechanical sieve shaker to ensure that each sample will be subjected to the same length of time and amplitude of shaking. Sieve brushes were used to clean the sieves before segregating the next sample. This is to ensure that earlier segregated samples do not contaminate the later samples as well as keep the sieves dry so as not to collate the sediments and interfere with the results of the segregation.
The sieving method has several limitations. The probability of a particle passing a given sieve in a given time of shaking depends upon the nature of the particle and properties of the sieve (Day, 1965). The dimensions of the particle might hinder the particle from passing through a given sieve opening, even though the particle has the correct mean size to pass though. Thus it will require extensive shaking before the particle is orientated into the “correct” orientation to pass through the sieve opening. Sieve openings are generally unequal in size, requiring extensive shaking before all particles have the opportunity of approaching the largest openings. A "complete" sieving can rarely be met in practical times of shaking, thus we decided to designate our own shaking time (15 minutes). We obtained the 15 minutes sieving time by performing a few trials sieving to see what is the minimum time required to achieve sufficient segregation of the samples without compromising the quality of the data.
As our data collection is discrete by nature, there is a need to interpolate the points to "fill in the gaps" as grain distribution is continuous by nature. Application of interpolation in the field of Geography can often be seen during the estimation of rainfall, temperature and vegetation species. There are several well known techniques around, including inverse distance weighted (IDW) interpolation and kriging.
IDW uses the idea of weighting each data by the inverse of the distance to the estimation location. This implies that the points further away from the estimation location will have an effect of only a power p of the inverse distance, thus not affecting as much as points that are nearer to the estimated point.
IDW assumes a homogeneous weight of influence by all the data points. Kriging, on the other hand, is preceded by a modeling of the spatial structure of the data; unlike IDW, kriging assumes that each data point has a different weight with respect to the estimated point. In addition, points that are behind another sample point will be "screened off" by lowering of weight, thus reducing the effect of this "blocked" point on the estimated point. This effect is known as the Screen effect (Wackernagel, 1995). Kriging estimates an underlying spatial variogram model to represent spatial variability. Furthermore, in kriging, as a statistical method, information such as the estimation of error can be known. This is obtained from the kriging standard deviation. With these considerations, we used kriging as our interpolation technique of choice. Our krigged raster layers of grain sizes were generated in ESRI’s ArcGIS.
IDW uses the idea of weighting each data by the inverse of the distance to the estimation location. This implies that the points further away from the estimation location will have an effect of only a power p of the inverse distance, thus not affecting as much as points that are nearer to the estimated point.
IDW assumes a homogeneous weight of influence by all the data points. Kriging, on the other hand, is preceded by a modeling of the spatial structure of the data; unlike IDW, kriging assumes that each data point has a different weight with respect to the estimated point. In addition, points that are behind another sample point will be "screened off" by lowering of weight, thus reducing the effect of this "blocked" point on the estimated point. This effect is known as the Screen effect (Wackernagel, 1995). Kriging estimates an underlying spatial variogram model to represent spatial variability. Furthermore, in kriging, as a statistical method, information such as the estimation of error can be known. This is obtained from the kriging standard deviation. With these considerations, we used kriging as our interpolation technique of choice. Our krigged raster layers of grain sizes were generated in ESRI’s ArcGIS.
Analysis
In every outthrust headland, in every curving beach,
in every grain of sand there is the story of the earth. - Rachel Carson
in every grain of sand there is the story of the earth. - Rachel Carson
Spatial regression
Variables
Statistical findings
We performed multiple regression analyses for each of the 9 grain sizes, to show how GIS can reveal possible spatial trends in the distribution each grain size. An ordinary least squares multiple regression was performed because it allows us to isolate the effect of one independent variable while controlling the effects of other variables.
There are several assumptions of multiple regression which are important when interpreting the regression coefficients (Mather and Openshaw, 1974). Firstly, effects of independent variables on grain size distribution are assumed homogeneous across the study area and marginal implicit variation of the grain size percentage is assumed equal. A change in the attributes will correspond to an absolute independent linear change in the percentage of grain size. Furthermore, an OLS regression assumes the error terms are independently and identically normally distributed with mean zero and constant variance, and not autocorrelated. The attributes cannot be multicollinear. While these assumptions may over-simplify the actual complex phenomenon of grain size distribution process, they may uncover trends at a general scale at a broader resolution.
There are several assumptions of multiple regression which are important when interpreting the regression coefficients (Mather and Openshaw, 1974). Firstly, effects of independent variables on grain size distribution are assumed homogeneous across the study area and marginal implicit variation of the grain size percentage is assumed equal. A change in the attributes will correspond to an absolute independent linear change in the percentage of grain size. Furthermore, an OLS regression assumes the error terms are independently and identically normally distributed with mean zero and constant variance, and not autocorrelated. The attributes cannot be multicollinear. While these assumptions may over-simplify the actual complex phenomenon of grain size distribution process, they may uncover trends at a general scale at a broader resolution.
Dependent Variable
- Grain size percentage
Independent Variables
- Nearest Distance to coastline (Cross-shore distribution)
- Nearest Distance to South-East shoreline (Cross-shore distribution)
- Nearest Distance to North-East shoreline (Long shore distribution)
These variables reveal possible topographical changes and general wave energy levels. They allow for spatial analysis to be done as the visualization of the spatial distribution of different grain size proportions will reflect the progression of the sand grains.
The cross shore distribution of grain sizes based on studies by Fox, Ladd and Martin (1966) and Komar (1977, 1998) is such that the "largest sand particles are generally located in the zone of the most intensive wave-breaking" (Komar, 1998), with the size of sand grains decreasing shoreward across the surf and swash zones. This is due finer grains requiring less energy to travel up shore as swash energy decreases with distance. The relation between energy and transportation rates can be seen from Figure.
Studies on the selective long shore transportation of sand grains by Evans (1939), Komar (1977) and Self (1977) also depict how because of the tendency for smaller sand grains to move up the beach face shoreward, they have a lower rate long shore transportation rate than larger grains which remain at the breaker zone and so are transported by the stronger long shore currents (Komar, 1977). Selective long shore sediment transport was found to be affected by fluvial currents from rivers by Self (1977), where the larger grains tend to be found nearer the river mouth due to reductions in current energy in these areas. The general direction of the long shore drift is shown to be going from east to west in that area according to Wong (1985).
The usage of these variables for measurement and statistical analysis is done to show how GIS can be used in studying the process of grain size sorting.
Other Independent Variables
- Within sand bar boundary (nominal)
- Percentage area covered by sea grass
These variables provide the controlling variables for this study, and can provide insights to the grain sizes which coincide at the locations. As the Chek Jawa is a unique area, there will be certain features that will influence the grain size sorting process as well or account for the local coastal morphology. There is the possibility that these variables may prove to be significant in accounting for the variations in grain size on the beach of Chek Jawa, adding to the influences on the grain size sorting process.
The coefficients significant at the 0.05 level are presented in a simplified, interactive manner in "toggle analysis" in the map view for visualization and "brushing" of the gradient of the distribution of the nine grain sizes. The bar-charts are normalized at the highest percentage and through the interactive mouseovers, the relative gradients can be discerned and grain sizes which vary at the highest/lowest rates can be discerned.
- Grain size percentage
Independent Variables
- Nearest Distance to coastline (Cross-shore distribution)
- Nearest Distance to South-East shoreline (Cross-shore distribution)
- Nearest Distance to North-East shoreline (Long shore distribution)
These variables reveal possible topographical changes and general wave energy levels. They allow for spatial analysis to be done as the visualization of the spatial distribution of different grain size proportions will reflect the progression of the sand grains.
The cross shore distribution of grain sizes based on studies by Fox, Ladd and Martin (1966) and Komar (1977, 1998) is such that the "largest sand particles are generally located in the zone of the most intensive wave-breaking" (Komar, 1998), with the size of sand grains decreasing shoreward across the surf and swash zones. This is due finer grains requiring less energy to travel up shore as swash energy decreases with distance. The relation between energy and transportation rates can be seen from Figure.
Studies on the selective long shore transportation of sand grains by Evans (1939), Komar (1977) and Self (1977) also depict how because of the tendency for smaller sand grains to move up the beach face shoreward, they have a lower rate long shore transportation rate than larger grains which remain at the breaker zone and so are transported by the stronger long shore currents (Komar, 1977). Selective long shore sediment transport was found to be affected by fluvial currents from rivers by Self (1977), where the larger grains tend to be found nearer the river mouth due to reductions in current energy in these areas. The general direction of the long shore drift is shown to be going from east to west in that area according to Wong (1985).
The usage of these variables for measurement and statistical analysis is done to show how GIS can be used in studying the process of grain size sorting.
Other Independent Variables
- Within sand bar boundary (nominal)
- Percentage area covered by sea grass
These variables provide the controlling variables for this study, and can provide insights to the grain sizes which coincide at the locations. As the Chek Jawa is a unique area, there will be certain features that will influence the grain size sorting process as well or account for the local coastal morphology. There is the possibility that these variables may prove to be significant in accounting for the variations in grain size on the beach of Chek Jawa, adding to the influences on the grain size sorting process.
The coefficients significant at the 0.05 level are presented in a simplified, interactive manner in "toggle analysis" in the map view for visualization and "brushing" of the gradient of the distribution of the nine grain sizes. The bar-charts are normalized at the highest percentage and through the interactive mouseovers, the relative gradients can be discerned and grain sizes which vary at the highest/lowest rates can be discerned.
The results of the statistical tests (to see, click "toggle analysis" in map view) show that relationships at 0.05 significance level exist between several of the independent variables and grain size percentages. Along with the visual representation, general aspects of the long shore drift flow may be suggested by the statistics; the beach at Chek Jawa seems to undergo selective long shore transportation. Cross-shore transportation also is accounted for and can be clearly seen. The impact of river flow can also be attributed by the distribution of the larger grain size spatial distribution as the larger particles are concentrated at the confluence of two channels and the river mouth, a unique situation in any case. So according to Komar (1998), variations in long shore sediment transportation patterns can be accounted for by four conditions or a combination of them. From using GIS, it can be generalized in this case that two of these situations appear to apply for the beach of Chek Jawa.
Results suggest that the area is one that is complex and subjected to many influences:
1. Velocity of less than 0.5 m/sec in this area (Wong, 1985) may prove to be significant as from the Hjulstrom curve would imply sediments between 0.01 mm and 0.15 mm may be affected due to sufficient energy of the tidal streams
2. Presence of river indicates source of sediments as well as a counteracting force to the long shore current
3. The sand bar may be the resultant feature of the tidal flow and the river flow due to the lowering of energy levels
4. Selective sorting may also have taken place as can be seen from the progressive reduction in grain sizes further away from the river where fluvial influence is the greatest (Self, 1977)
Results suggest that the area is one that is complex and subjected to many influences:
1. Velocity of less than 0.5 m/sec in this area (Wong, 1985) may prove to be significant as from the Hjulstrom curve would imply sediments between 0.01 mm and 0.15 mm may be affected due to sufficient energy of the tidal streams
2. Presence of river indicates source of sediments as well as a counteracting force to the long shore current
3. The sand bar may be the resultant feature of the tidal flow and the river flow due to the lowering of energy levels
4. Selective sorting may also have taken place as can be seen from the progressive reduction in grain sizes further away from the river where fluvial influence is the greatest (Self, 1977)
Discussion
You can tell all you need to about a society from how it treats animals and beaches. - Frank Deford
Applicability of GIS
Limitations
Extensibility
Bibliography
The application of GIS to this study has simply made this study on the process of grain size sorting at Chek Jawa more easily understood and can become a possible paradigm for other grain size sorting studies. With the inclusion of an interactive and user-friendly interface, such a non-proprietary web-based GIS may encourage learning through data exploration (Slocum, McMaster, Kessler and Howard, 2005). The geographic visualization of the spatial distribution of the grain sizes through thematic brushing allows patterns to be easily observed and may spur further inquiries in understanding the possible influences on this phenomenon in a local setting. Using Chek Jawa as an example for lessons in coastal environments creates a local reference point for students to identify and associate with, in contrast to remote examples in foreign contexts.
Understanding the grain size sorting process primarily remains within the academia domain where the beach processes still intrigue academics. But with GIS, there is an increased extensibility of such studies to the planning of management practices of coastal areas. Besides spatial analysis, the temporal dimension can be included as well. This can further lead to strategic planning and the construction of prediction models.
The foundation for the use of GIS in the study of one of the most inexplicable nearshore process has been laid and its applicability has been introduced. Further studies could include continuation in spatial and temporal data development.
Understanding the grain size sorting process primarily remains within the academia domain where the beach processes still intrigue academics. But with GIS, there is an increased extensibility of such studies to the planning of management practices of coastal areas. Besides spatial analysis, the temporal dimension can be included as well. This can further lead to strategic planning and the construction of prediction models.
The foundation for the use of GIS in the study of one of the most inexplicable nearshore process has been laid and its applicability has been introduced. Further studies could include continuation in spatial and temporal data development.
There are many constraints and limitations to this project at its current stage. Computers only perform calculations for us mechanistically. Amongst the most obvious would be that the accuracy of the samples and the lack of in depth considerations of the multiple influences in such a complex, dynamic beach environment. The sampling methods can be said to be inadequate to fully map out the gradation of sand grains for long shore drift studies. Important factors that affect the grain size sorting like the elevation and topology of the area, wave intensity levels and wind data were not taken in consideration given the scope and time constraints of this project. All these are also affected by the lack of data over the temporal dimension.
With such limitations, it is not possible to accurately and comprehensively speculate about coastal processes based on the material from this study. However, despite issues with data resolution, while geostatistical methods like kriging masks local variability at finer resolutions, they can convey information about general trends. Once again, the main purpose of this study is to illustrate the applicability of GIS in undertaking such studies on beach processes, and to foster sentiment towards natural processes in public viewers. We have approached this study from a general perspective, considering time and logistical constraints.
With such limitations, it is not possible to accurately and comprehensively speculate about coastal processes based on the material from this study. However, despite issues with data resolution, while geostatistical methods like kriging masks local variability at finer resolutions, they can convey information about general trends. Once again, the main purpose of this study is to illustrate the applicability of GIS in undertaking such studies on beach processes, and to foster sentiment towards natural processes in public viewers. We have approached this study from a general perspective, considering time and logistical constraints.
A raster dataset, as a two-dimensional array of values, is platform-independent and can be conveniently superimposed on other georeferenced maps as overlays for further analyses; for example, overlaying marine biological data patterns on grain size layers using GIS can help in uncovering possible correlations between organisms and substrate quality.
In the pedagogic realm, these thematic overlays can be used as an effective teaching tool in communicating such phenomenon.
This case can serve as a starting point which can be expanded to the other beaches in Singapore, especially in context of the heavy modifications to the coastline in recent years. Knowing how sand grains are sorted can possibly give an idea of the energy level of a particular beach environment (Nordstrom, 1977). This will have implications on the methods of beach protection employed. For instance, predicting the possible morphological changes to the beach should beach nourishment activities take place will help in determining the effectiveness of the activities.
This case can serve as a starting point which can be expanded to the other beaches in Singapore, especially in context of the heavy modifications to the coastline in recent years. Knowing how sand grains are sorted can possibly give an idea of the energy level of a particular beach environment (Nordstrom, 1977). This will have implications on the methods of beach protection employed. For instance, predicting the possible morphological changes to the beach should beach nourishment activities take place will help in determining the effectiveness of the activities.
Bird, E. C. F. (2000) Coastal Geomorphology: An Introduction. John Wiley & Sons. Chichester, New York. pp. 95-98.
Christman, M. C. (2000) A Review of Quadrat-Based Sampling of Rare, Geographically Clustered Populations. Journal of Agricultural, Biological, and Environmental Statistics, Vol. 5, No. 2, pp. 168-201.
Cochran, W.G. (1953) Sampling Techniques. Johan Wiley and Sons, Inc., New York.
Evans, O. F. (1939) Sorting and Transportation of material in swash and backwash. Journal of Sedimentation Petrology, vol. 9, p.28-31
Loh, K. S. (2008) Mass mortality and recruitment of intertidal marine invertebrates in Chek Jawa, Singapore. Undergraduate Research Opportunities in Science project report submitted to the Department of Biological Sciences, The National University o
Christman, M. C. (2000) A Review of Quadrat-Based Sampling of Rare, Geographically Clustered Populations. Journal of Agricultural, Biological, and Environmental Statistics, Vol. 5, No. 2, pp. 168-201.
Cochran, W.G. (1953) Sampling Techniques. Johan Wiley and Sons, Inc., New York.
Evans, O. F. (1939) Sorting and Transportation of material in swash and backwash. Journal of Sedimentation Petrology, vol. 9, p.28-31
Loh, K. S. (2008) Mass mortality and recruitment of intertidal marine invertebrates in Chek Jawa, Singapore. Undergraduate Research Opportunities in Science project report submitted to the Department of Biological Sciences, The National University o