Köppen Climate Classification
A GIS Model Using PRISM Data
Introduction
One of the basic elements of the environment that influences all life is air. Weather is the condition of the air at a given moment and location. Climate, however is the total occurrence of weather at a given location over some extended period of time. Weather can be easily observed, measured, and stated. Climate cannot be evaluated exactly because it "...represents a complex and abstract idea which has no concrete existence at a given instant." (Blair, 1942). Climate exerts an undeniable influence on the surface of the earth. Climatic gradients influence the spatial distribution of humans, animals, vegetation, and microorganisms. Within the human spatial distribution, occupation, architecture, economy, health, and the general manner of living are all influenced to a certain degree by the climate. Additionally, the foraging and mating habits of animals, the growth rate and density of vegetation cover, and the abundance of microorganisms in a given location are affected by the climate. This phenomena that influences where we live, what we do, how we do it, what we do it with, and generally all aspects of our everyday lives deserves considerable attention. It is through a better understanding of climate that we may increase production and output in a number of industries and insure healthy and stable ecological and biological productivity.
A climate classification is an attempt to divide up a predefined area into regions (or zones) with an approximately homogeneous set of climatic conditions. A zone is an area in which one type of climate dominates and where the major climatic elements are much the same. It is not an area of complete climatic uniformity. The boundaries between climatic regions should not be thought of as distinct lines but rather transition zones (i.e., belts in which the climate changes gradually from one type to the next). There is substantial room for difference of opinion and practice as to where a transition zones may occur. Ultimately, there is no exact number of climate zones that any given area is required to have. Strahler (1951) notes that success is only partially achieved when "...the scientist attempts to devise schemes of classification that will include all variations in climate, yet permit them to be placed in several clearly defined and easily distinguished groups."
The importance of developing a climate classification was realized in the early part of twentieth century. In 1918, Wladimir Köppen (1846-1940) developed perhaps the best known and most widely recognized climate classification of the world. This classification by Köppen divided the land areas of the world into five major climatic categories largely based on temperature and precipitation (Blair, 1942). Over the years, the original Köppen classification has undergone some modifications, e.g. the addition of a sixth major class. The Köppen classification represents one attempt to understand, classify, and spatially represent areas of similar climatic condition.
In order to produce the Köppen classification within a geographic information system (GIS), gridded estimates of monthly and annual precipitation and temperature data and a digital elevation model are required. This data was acquired from the Oregon Climate Service web site. PRISM (Parameter-elevation Regressions on Independent Slopes Model) is a model that uses point data and a digital elevation model (DEM) to generate gridded estimates of climate Parameters (Daly et al., 1994). The model was designed and tested using precipitation, temperature, and DEM data from the PRISM project.
Project Goal
The goal of this project was to write an AML in ARC/INFO that would produce a climate classification for a given area of interest based on the Köppen Climate Classification criteria.
Model Description
A flow chart, shown below, displays the overall model. The main AML for the model is Köppen.aml. Six other sub-AMLs, contained in a directory named tools, run from Köppen.aml in order to produce the final classification grids.
Program Description
In order to run the run the classification for a given area, Köppen.aml and the directory tools must be copied into a workspace. Köppen.aml is the main AML for the classification and is run from the "ARC:" prompt. The "tools" directory contains 6 sub-AMLs that run from Köppen.aml. They are as follows:
The user is then presented with a popup window, shown below, containing information about the AML.
The user must copy the needed files, listed in the popup window, into the newly created data directory.
Input.aml is run and prompts the user with a series of 29 menus; see examples below. The first thirteen menus ask the user to enter the ASCII mean annual precipitation files for each month of the year and annually. These files must have a .asc extension.
The following thirteen menus ask the user to enter the ASCII mean average annual temperature files for each month of the year and annually. These files must also have a .asc extension. The next menu to appear asks the user to enter a DEM. The DEM must have a .dem extension. Next, a menu appears asking the user to enter a polygon clip coverage of the area of interest. Finally, a menu appears asking the user to enter a projection file. The projection file must have a .prj extension. The polygon clip coverage and the projection file must have the same output projection. The input projection for the projection file must be 'geographic' with units in decimal degrees ('dd') if the user is working with PRISM data. At that point, the user input to the model is concluded.
Asciigrid.aml is run and takes the precipitation, temperature and DEM files and converts them into ARC/INFO grids. Projection.aml is run and projects the grids based on the input projection file. Latticeclip.aml is run and the grids are clipped with the input boundary coverage. Grids.aml is run and carries out, mainly through grid algebra statements, the calculations needed to produce subgrids for the Köppen classes. Classes.aml is run and the Köppen classes are defined through a series of "if" and "do" statements. Ultimately, the final grids are produced.
In summary, to run the model, the steps outlined below would be followed:
Results
The model was run for three areas of interest: the entire United States, the northwest states of Washington, Oregon, and Idaho, and the state of Idaho. For the entire United States, the model takes about 35-40 minutes to run on an NT Workstation with a 266Mz processor and 128 megabytes of RAM. There are no known bugs in the model. The model produces individual grids of each class that exist for the given area of interest. The grids that were produced for the three areas of interest mentioned above match up well with the original Köppen maps. In particular, the major classes of the United States match nicely (see Critchfield, 1983). Therefore, the results of this project affirm the accuracy of the PRISM data.
References
Blair, T. A., 1942: Climatology. Prentice-Hall, Inc., 484.
Critchfield, H. J., 1983: General Climatology. 4th ed. Prentice Hall, 453 pp.
Daly, C., R. P. Neilson, and D. L. Phillips. 1994: A statistical-topographic model for mapping climatological precipitation over mountainous terrain. Journal of Applied Meteorology, 33, 140-158.
PRISM, cited 1998: Climate Mapping with PRISM. Available on-line at: PRISM
Strahler, A. N., 1951: Physical Geography. John Wiley and Sons, Inc., 733.
