Method


BC Basemap and Geocoding

A basemap of B.C. was needed to generate precipitation maps from the precipitation data we obtained from Dr. Brian Klinkenberg. The UBC Geography Lab database has a shapefile of the border of B.C., which is an open polygon, a line feature. We then closed the border to incorporate Queen Charlotte Island and Vancouver Island, by joining the end points of the border line. The border shapefile was then converted to a coverage, which we ran the topology clean command to build polygons of B.C. land mass. We then proceeded to geocoding and started with an excel file with the data of 1723 climate stations in BC. After minor adjustment to the data, we saved the sets of data into .dbf (dBASE IV) files, and converted its projection in ArcCatalog. First, we clicked on the Create Feature Class > from XY table. In the panel, we selected the X field as the longitude, Y field as the latitude, Z field as the elevation. We then clicked on the Spatial Reference of Input Coordinates, and select the geographic coordinate system > North America Datum 1983. Next, we clicked on the Advanced Geometry Options, selected the “use a different spatial reference”, and imported a shapefile with a projected coordinate system of Albers. After converting these 2 sets of data, we added the files into the ArcMap.
Climate stations in BC
Fig.1 - Geocoded climate station in BC
(click on map to enlarge the size)

Distance from Coast

To describe the continentality influence on precipitation pattern in B.C., we separated the B.C. land mass into 5 classes with decreasing coastal influence away from the coast. First, an approximate west coastline of North America along the B.C. area was drawn in. The coastline incorporated Vancouver Island as part of the B.C. mainland, based on the assumption that Georgia Strait would not have a significant influence on the precipitation east of it, and also that the Insular Mountains of Vancouver Island would act as a barrier to the moisture traveling inland from Pacific Ocean. Queen Charlotte Island was separated from B.C. mainland due to its distance from it, and another coastline was drawn in to approximate the west coast of Queen Charlotte Island. Buffers of 200 km each were put in for both coastlines to classify the continentality effect. The B.C. mainland and Queen Charlotte Island buffer shapefiles were then combined and converted to a raster. Using raster calculator and the DEM of B.C., the buffer was clipped to show the zones for B.C. land mass only.
Distance from Pacific Coast
Fig.2 - Continentality Effect: Distance from the Pacific coast
(click on map to enlarge the size)

Valley System

An initial micro-valley system map was generated. Using the DEM of BC a slope calculation and a neighbour statistics of elevation were performed to generate a mean elevation with neighbouring 15 raster cells. Taking the difference of the orginal elevation and the mean neighbour elevation will generate area of valley bottom with negative values. Combining these area with a slope of less than 5 degrees through a raster calculation will generate valley bottom area. Taking into consideration that lakes will not have negative raster value for elevation differences, a lake vector file is incorporated into map.
Micro-Valley System
Fig.3 - Valley System at a small scale
(click on map to enlarge the size)

However upon looking at the micro-valley system map, it did not incorporate the Northern Interior Plateau. At a larger, macro scale valley system, the Interior Plateau constitutes an important valley system and need to be considered. After struggling with exploring method to classify such area, it was decided that the area will be selected through the BEC Zone classification. Since BEC Zones have distinct climatic regime and the vegetation grown is the result of temperature and moisture, I argued that combining certain BEC zones can approximate the valley system. These zones are BG, PP, IDF, ICH, MS, SWB, SBS, MH. Upon selecting these zones, it yield an more accurate depiction of where valley system are beside it included most micro-valleys and the plateau region. A conversion of vector features (BEC zones) to a raster image was performed for the generation of the new precipitation model.
Macro-Valley System
Fig.4 - Valley System at a large scale
(click on map to enlarge the size)

Elevation and Aspect

The elevation map is quite simply the DEM obtain from the database of the Geography computer lab, however reclassification are needed for the new precipitation model generation.
Digital Elevation Model
Fig.5 - DEM of BC showning elevation
(click on map to enlarge the size)
As for creating an aspect map a simply DEM surface analysis was performed in 769m cell resolution. Then, we produced an aspect map of 769 m resolution by clicking on the Spatial Analysis > Surface Analysis > Aspect. Next, we classified the values into 6 classes by selecting the Spatial Analyst > Reclassify. In the panel, we selected the “manual” method, 6 classes, and gave the breaks values as 0, 45, 135, 225, 315 and 360. 315 to 45 is North Aspect, 45 to 135 is East Aspect, 135 to 225 is South Aspect, and 225 to 315 is West Aspect. Since North and South Aspect does not effect precipitation too significantly, the value of 0 is assigned and we did not use it in our analysis. The new values of 1 and 2 represented east facing and west facing respectively. For the display purpose, we also generated a classified aspect map in a 4km resolution.
East-West Aspect Map
Fig.6 - West/East facing aspect map showing the windward and leeward side of the mountain respectively
(click on map to enlarge the size)

Precipitation Maps

Precipitation maps were generated from the precipitation data obtained from 1723 climate stations in B.C.. For the purpose of comparison, the daily precipitation data were classified into 4 seasons, with 2 different classifications of seasons:
  1. 3 months seasons:
  2. Solstice and equinox seasons (http://www.crh.noaa.gov/ind/seasons.txt) – pick the 2004 solstice from the website:
Five maps were produced: an average annual precipitation map, and four others each for one season. The maps were generated using Ordinary Kriging as the interpolation method over all the station points. The extent of the Kriging output maps were set to cover the entire B.C., and then exported to raster. Using raster calculator and the DEM of B.C., the B.C. land area was clipped from the raster layers. These raster layers would serve as the precipitation maps of B.C. showing the different precipitation pattern in a year and in four seasons. And the average annual precipitation map would be the model with which our new precipitation model will be compared.
A.B.
December 21st to March 19th March 20th to June 19th
C.D.
June 20th to September 21st September 22nd to December 20th
Fig.7 - Seasonal daily average precipitation A.Winter Solstice - December 21st to March 19th, B.Vernal Equinox - March 20th to June 19th, C.Summer Solstice - June 20th to September 21st, D.Autumnal Equinox - September 22nd to December 20th.
Annual Daily Average Precipitation
Fig.8 - Annual Daily Average Precipitation in BC
(click on map to enlarge the size)

New Precipitation Model

After all the layers have been properly converted into raster format, a simple raster calculation was performed to combine the 4 factors we considered as important for precipitation pattern. The weight of each variable was determined relatively. For example the Coast have the most influence on precipitation so the weight for 200km from coast have a high weight. Also in the interior where continentality is fairly constant, valley system is a key determinant. At the Rockies continentality is not as much of a factor as elevation.
New Precipitation Model
Fig.9 - A combination of elevation, aspect, distance from coast and valley system factors to generate a more accurate model for precipitation pattern in BC
(click on map to enlarge the size)






BACK    NEXT