Introduction

 

Data Collection and Preparation

According to the Environmental Assessment Office's Salmon Aquaculture Review (1997), salmon farm has the following biophysical siting criteria:

Factor Requirement
Temperature - Minimum > 0°C
Temperature - Maximum < 25°C
Salinity 0 - 34 ppl
Depth - Minimum 5 m
Wave height - Maximum 3 m
Wind speed - Maximum 100 knots
pH 6.5 - 8.2
Dissolved Oxygen > 4 mg/L
Current speed - Minimum 5 cm/s
Current speed - Maximum 140 cm/s

The above criteria are not all presented in the project. Wind speed, for example, is far from the maximum. Therefore it is not included in the actual analysis after the data are examined.

The data are collected from various environmental organizations, in comma separated tabular form. The sources of data including the Department of Fisheries and Ocean (DFO Site 1 and DFO Site 2), National Oceanographic Data Center (NODC), and the SIS Lab for the base map.

Wave information:
The single file is combined from a collection of comma separated value files from DFO Site 1, imported into Microsoft Excel . Each file contains a time series of one location. Information included are station ID, date (date and time), quality, latitude, longitude, depth, characteristic significant wave height (m), characteristic significant wave height (reported by the buoy) (m), maximum zero crossing wave height (reported by the buoy) (m), wave spectrum peak period (s), wave spectrum peak period (reported by the buoy) (s), direction from which the wind is blowing, horizontal wind speed (m/s), horizontal scalar wind speed (m/s), gust wind speed (m/s), atmospheric pressure at sea level (mb), dry bulb temperature (°C), and sea surface temperature (°C) (Sample file). As such, various data are derived these files.

Each file has been reduced to contain only information after year 2001. Two entries were selected from each file, where possible: one on January 1, 2001 (at 12:30pm), and one on July 1, 2001 (at 12:30pm) to represent a sample winter and summer data. In case where no such data exist, the closest date (and time) was picked. The winter and summer sample were put into different spreadsheets and subsequently exported to DBASE IV format (Figure 1). This set of files derived depth, sea surface temperature (SSTP), wave height, and acidity (pH).

Figure 1

Salinity:
Salinity is derived from a combination of lighthouse text files from DFO Site 2. Each file is imported into excel, and contains a time series detailed to average per month. The file is exported to DBASE IV format but is not distinguished into a winter file and summer file, as it is identified in the columns (Figure 2).

Figure 2

Acidity (pH):
Acidity data is obtained from an archival site from NODC. Similar to the previous factors, the data is divided into a winter set and the summer set (due to the format of the data, it is more convenient to separate them into two files) (Figure 3).

Figure 3

Rasterization:
Study area: Two study areas were drawn on the base map of GCCS059b. Study_area.shp was a larger version that includes an area from 46°N to 58°N, and 122°W to 139°W. The purpose of this is to include all of the area covered by the available data points. This is recognition of the problem with interpolation, that the points lying just outside the boundary is delineated but could otherwise have influence on the interpolated surface. A New_study_area.shp was created that covers the actual interested study area, including from 48°N to 56°N, 122.5°W to 134°W. The New_study_area.shp is used in the later procedures to clip the layers into a right size.

The Study_area.shp is 739 x 1050 (grid cell size = 0.017989 dg) in terms of raster cells. This is the bases for all the subsequent interpolations.

All the data remains unprojected, because distance is not involved.

Since each factor is in tabular form, they have to be interpolated. Each data table is imported into ArcView, and added as an event theme since each data entries have a coordinate attached.

Base map: Taken from ArcInfo coverage GCCS059b, it was clipped using the New_study_area.shp, then merged based on attributes. The attribute Pruid was used because it has a uniform value and I only need the base map as a constraint (i.e. obviously these factors does not apply to on-land closed containment systems). The purpose of the base map is to act as a mask to the final evaluation (Figure 4).
Data Collection
Methodology
Spatial Analysis

Interpolation

Overlay
Fuzzy
MCE
Result
Discussion
Conclusion
 
 
Email me
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 


Figure 4

 

Depth: Surface created using depth information from wave_height_information_summer_2001.dbf. IDW (nearest neighbor, neighbor = 4, power = 3) was used instead of spline because spline creates a surface that is always larger than 0 (depth are interpreted as -(meters)), while IDW does not give a positive elevation. The number of neighbor is chosen to be 4 because the points in reality are not as clustered as it now appears. There are a few points scattered along one side of the Queen Charlotte Island and another few points scattered on the other side. Therefore if a large number of neighbor is chosen, it can be expected that the points that are separated by physical features will have influence on the other. A power of 3 is given to further reduce the significance of points far away (as in IDW, the larger the power, the less further point influences). As a result, despite it is more eye pleasing with spline than with IDW, the method adopted is still IDW.

Salinity: Surface created using salinity information from salinity retyped.dbf. Tension spline was used to create a smooth surface (as opposed to the often jaggy surface created with IDW), with weight = 1, number of points = 4. The reason tension spline was used over the regularized spline is the surface values are closer constrained by the sample data. The combination of weight = 1 and number of points = 4 is a result of trial and error, and this creating the "best" surface (i.e. limited extreme values).

Acidity (pH): Surface is created using pH_information.dbf. Tension spline was used again with the same parameter (weight = 1, number of points = 4) for the same reason as salinity. Tension spline is a general-purpose surface interpolator and thus good for this application.

Sea Surface Temperature (SSTP): Two surfaces are created for the different season. Tension spline was used again, because the data set is derived from the same set of information containing depth. A winter set and a summer set was created using the respective wave_height_information file.

Wave height (wvht): A winter and a summer surface are created using the two sets of wave_height_information files, also using tension spline. Different to other layers, the weight is set to 4.5 and number of points is set to 7. This is done thru trial and error to minimize the area with a negative wave height.

Clipping: The original data layers are all interpolated to a larger extend than needed. Clipping the layers to the boundary delineated by new_study_area.shp requires the use of an ArcScript downloaded from ESRI website. This is because the geoprocessing function in ArcView does not support clipping of non-shapefiles. Clipped GRID layers are exported in binary raster format for import in Idrisi.

back | home | next