CHAPTER II: BACKGROUND AND LITERATURE REVIEW

This chapter discusses the general climate, natural vegetation, and agriculture patterns in Idaho. The state's rugged, highly diverse terrain varies between steep mountain ranges and rolling prairie land across forty-four counties (Figure 1). This muliform landscape, coupled with the maritime influence of the Pacific Ocean some 640 kilometers to the west, creates a complex climatic environment for the state's 22,000 farms and ranches (Idaho Agricultural Statistics Service, 1998). Additionally, a review of literature relevant to this thesis is discussed.

Climate of Idaho

Since this thesis attempts to divide Idaho into a relatively small number of agroclimate zones, a background into the general climate of the state is necessary. This section provides generalities about the climates of Idaho.

The wide range of elevation and complex topography of Idaho combine to influence its climate and produce a varied climate (Tables 1 and 2). Numerous steep, jagged peaks and ridges that project from the main mountain ranges eventually give way to rolling prairie land, high mountain valleys, deep canyons, and high deserts. This varied landscape takes place on approximately 216,000 square kilometers between an elevation of 225 meters, the confluence of the Clearwater and Snake Rivers near Lewiston, and 3860 meters, Borah Peak. A considerable amount of area, approximately 165,000 – 170,000 km² lies between 760 and 1830 meters above sea level (USDA, 1941). The Snake River Plain in southern Idaho is the largest expanse of relatively flat topography totaling approximately 40,000 – 50,000 km². It increases in elevation from 556 meters at Weiser, at the extreme western edge, to 1524 meters in the extreme east, near Yellowstone Park (USDA, 1941).

Table 2. NWS station temperature and precipitation extremes for 1961-1990.

The climate is milder than its latitude would indicate due to the influence of the Pacific Ocean. Prevailing westerly winds originating in the Pacific Ocean some 640 kilometers to the west influence both the summer and winter temperatures with a stronger maritime effect on the latter. This influence is felt more in central and northern Idaho than south of the Salmon River. The climate in eastern Idaho (the central mountains and areas near Yellowstone National Park) is more continental in character than in western and northern Idaho resulting in a greater range between winter and summer temperatures. In winter, the strong winds, particularly in the northern part of the state, result in a maritime influence that helps to produce relatively mild conditions. However, on some occasions dry arctic air from Canada spills west of the continental divide producing clear skies and a distinct drop in temperatures. One such arctic outbreak in December 1968 resulted in record cold temperatures over much of the state with the minimum temperature reaching - 42° C at Moscow (Abramovich et al., 1998). The lower Snake and Salmon River valleys near Lewiston and Riggins have the warmest mean minimum nighttime winter temperatures whereas the high mountain valleys around areas such as Stanley have the lowest.

In summer, rainfall, cloud cover, and relative humidity are at their annual minimum due to a weakening of the westerly winds which allows continental climatic conditions to prevail (Abramovich et al., 1998). The coolest mean summer temperatures occur in the high mountain valleys near Stanley while the warmest summer temperatures occur in the Snake River Plain southwest of Boise.

The maritime air from the prevailing westerly winds is Idaho's major moisture source. This maritime influence is strongest in northern Idaho. Weather systems heading east off the Pacific coast encounter north-south mountain ranges and lower, relatively flat areas from the coast of Washington to northern Idaho. These mountain ranges force much of the precipitation carried by the eastward-moving winds to be expunged before reaching Idaho. The Olympic Mountains in western Washington see much greater precipitation amounts than those in the Puget Sound area; the leeward side of the mountain range. The Cascade Mountains of Washington produce a precipitation rise again, with decreasing amounts recorded in the Columbia Basin. The Bitterroot Mountains of Idaho raise the annual average once again with lesser amounts seen to the east in Montana. These successive mountain ranges illustrate the influence topography has on precipitation. In general, for Idaho, altitude is a more important factor of control than latitude as far as temperature and precipitation are concerned.

Annual average precipitation in Idaho ranges from about 175 mm near the towns of Grand View and Bruneau in Owhyee county to over 2000 mm along the Idaho-Montana border in Bonner, Clearwater, and Shoshone counties. In most areas of the state, December and January have the highest monthly precipitation. Monthly totals decrease irregularly until July and August, which are the driest months. All areas of the state have dry summers and no measurable precipitation in August is fairly common. After August, monthly amounts increase until January. With the exception of southeastern Idaho, most areas have a distinct wet winter-dry summer precipitation pattern. In southeastern Idaho (from Salmon to Idaho Falls to Grace) two precipitation peaks are seen: one in late fall and the other in late spring (Abramovich et al., 1998).

The average annual snowfall ranges from less than 0.3 meters around Lewiston, in Nez Perce county and Swan Falls, in Ada county, to about 20 meters along the Idaho-Montana border in Idaho and Shoshone counties. It exceeds 1.5 meters along the Idaho-Wyoming border and the northern and northeastern parts of the Panhandle, and exceeds 10 meters in northern Fremont county and the northwestern parts of Adams and Valley counties. Along the Snake River Valley from southern Bingham county to southern Washington county, the average annual amounts can range from less than 1 meter to about 2 meters. Snow has been recorded

at some stations in the state in every month of the year (Abramovich et al., 1998).

The last killing frost in the spring usually occurs in early to mid-May over the western Snake River Valley near Nampa and Caldwell and in the areas surrounding Lewiston. In the higher elevations in Lemhi, Idaho, and Custer counties the average date is around mid-July. The first killing frost in the fall usually occurs around early to mid-August in the central mountains. Around Lewiston and in areas south of Boise, it is usually deferred until October. Over much of the remainder of the state it is generally recorded some time in September. The growing season varies greatly throughout the state and the year-to-year variations can be considerable. The average length of the growing season ranges from less than 20 days in the higher elevations of Lemhi, Idaho, and Custer counties to about 200 days in the immediate vicinity of Lewiston. The central Snake, lower Boise, Payette, and Weiser River Basins experience approximately 150 or more days (NOAA, 1977). Over much of the Panhandle it exceeds 100 days, and in some localities in these areas even 130 days. Some of the high valleys around Stanley see freezing temperatures in every month and the land is predominately used for grazing.

Statewide, there is an average of 167 clear days during the year, ranging from 21 in July to 9 in January. There is an average of 94 partly cloudy days during the year, ranging from 10 in May to 7 during the late fall and winter. Finally, 104 cloudy days a year are seen on average, ranging from 15 in January to 3 each in July and August (USDA, 1941).

Natural Vegetation and Agriculture in Idaho

A great deal of the variation seen in climate correlates well with differences in natural vegetation due to the significant influence that climate exerts on vegetation. In fact, such words as desert and tropical are often applied to both climate and vegetation (Vankat, 1979). If a discussion of climate is undertaken, a discussion of vegetation should follow. Therefore, both the natural and introduced (i.e., agricultural) vegetation patterns of Idaho are discussed in this section.

Vegetation is a given combination of life forms and competing taxa with relatively uniform ecological requirements that dominates much of the appearance of the world’s landscapes and greatly influences human activities in many areas. Plants are stationary objects that possess distinct physical properties. "Real vegetation" includes all types of vegetation present at a time of observation. Therefore, this term includes man-introduced vegetation types such as agricultural crops. The term "natural vegetation" is regarded as vegetation that develops without appreciable interference or modification by man. In many locations, Idaho's land surface does not bear natural vegetation due to activities such as ranching, agriculture, forest thinning, and urbanization.

"Climax vegetation" in the conterminous United States is often thought of as vegetation that existed at the time when European settlers first appeared in the 16^th to 18^th centuries. The notion of white-man being the first to influence the vegetation of North America is challenged by some because, to a certain extent, the Native Americans exercised a disrupting influence. Therefore, an exact date when vegetation was considered pristine is somewhat a matter of opinion. The question of timing is raised since the dates of settlement appeared in different locations at different times. To circumvent this problem, the term "potential natural vegetation" was developed and is defined as

Man’s activities and influences prior to mapping are permitted to stand while future climatic fluctuations are eliminated. Therefore, "potential natural vegetation" implies a specific period of time and must bear a date if it is to be considered meaningful. Climax vegetation as it was when the first man ever beheld it occurred so long ago that it may have little or nothing in common with the vegetation of the 20^th century (Küchler, 1964).

Current vegetation in Idaho ranges from coniferous forests in the mountains to high scrub brush desert in the Snake Plain to grasses in the Palouse and other areas. This wide range of vegetation types results in economic activities that range from tourism to logging to agriculture. The majority of production agriculture takes place in the Palouse and the Snake River Plain although agriculture is found in each county with the exception of Shoshone (Figure 2). Additionally, gardening and orchard operations are found throughout the state. Idaho’s economy is strongly influenced by the forest products industry. Idaho ranks among the seven largest producers of softwood lumber in the United States producing 5 percent of that material for the nation (Idaho Department of Commerce, 1999). This plentiful natural resource has resulted in many building materials companies establishing businesses in the state. Over 200 companies make up the forest products industry in Idaho (Idaho Department of Commerce, 1999). Major paper and wood product manufactures rely on Idaho’s forest resources for their industry.

Table 3. Idaho's rank in the nation's agriculture.

Agriculture and food processing is Idaho’s largest industry, contributing 25% to the Gross State Product (Idaho Agricultural Statistics Service, 1998). Farming and agriculture-related business accounts for cash receipts totaling almost $3 billion and Idaho employs over 18,000 persons in food processing operations and more than 32,000 work on farms and ranches (Idaho Department of Commerce, 1999). There are approximately 22,000 farms and ranches in the state totaling about 55,000 km² with an average farm and ranch size of 2.5 km² (Idaho Agricultural Statistics Service, 1998). The state ranks among the top 5 U.S. producers of 14 agricultural commodities (Table 3). Major national food and beverage companies rely on Idaho's agricultural commodities as their source of supply as well as to maintain their reputations for high quality products. Idaho’s agriculture is as diverse as its climate ranging from traditional crops such as potatoes and grains, to specialty crops such as fruits, mint, and Christmas trees.

Literature Review

There are many different climate classifications and no one classification can be all things to all people. They use various combinations of climatic variables and different threshold values to delineate boundaries. Many classifications have a similar goal, which is to accurately map the climate of a large area based on a relatively few number of point observations. Not all classifications are necessarily developed for agricultural purposes. This does not necessarily mean that classifications developed for other purposes cannot have some agricultural applications. This section discusses classifications that range from relatively general and worldwide to those specifically developed for agriculture and encompass geographic areas as small as the Pacific Northwest.

A climate classification should do four things: (1) collate the vast amount of data into a manageable form, (2) be easy to apply, (3) be based on measurable meteorological variables, and (4) have well-defined objectives. It became possible to classify climates worldwide when actual observations covering much of the world's land surface were available in the early 20^th century. As large amounts of weather data began to be compiled, annual and monthly temperature and precipitation maps were created. When these maps were superimposed on one another, interesting patterns became apparent. These patterns were first discussed among botanists who began to notice a close, but not exact relationship between temperature and precipitation and plant life. The study of this relationship is known as phenology and is formally defined as the study of periodic biological phenomena, such as flowering, breeding, and migration, as related to climate. Expanding on the early approaches by botanist, climatologists such as Wladimir Köppen (1846-1940) developed climate classifications that were strongly influenced by the distribution of plant species.

Two techniques, or approaches, are generally used to create a climate classification: geographical techniques and statistical techniques. Geographical techniques, which were more prevalent in the early part of this century, involve the manual preparation of climate maps through topographic analysis. Statistical techniques, which have become increasingly popular in the last thirty years with the advances made in the computer industry, include methods such as distance-weighting algorithms and multivariate analyses.

Arguably, the best known and most widely recognized climate classification of the world was developed by Wladimir Köppen in the early 1900s. Köppen worked out the first drafts of the system as early as 1900 and 1918 and summarized the history and argumentation of the system in 1936 (Lamb, 1972). This classification by Köppen divided the land areas of the world into five major climatic categories based largely on temperature and precipitation (Blair, 1942). The major, or 1^st order zones, are designated by capital letters: A, B, C, D, or E. A, C, D, and E are based upon temperature while B has a precipitation meaning. Zone A represents tropical climates, Zone B denotes dry climates, Zone C depicts warm, temperate rainy climates with mild winters, Zone D represents cold forest climates with severe winters, and Zone E designates polar climates (Trewartha, 1954). A, C, and D are all essentially tree climates, B is a steppe grassland/desert climate, and E is a tundra/polar desert climate (Lamb, 1972). Within each major zone, 2^nd and 3^rd order subgroups are delineated. As with any climate classification, two parts are necessary to gain insight into a particular location: a map displaying the zones and a table describing characteristics about each zone. If one were to look at Köppen’s hand-drawn map and see the Zone ‘Cfb’, the Köppen table (an excerpt is shown in Table 4) would describe the characteristics associated with the zone as:

Table 4. Characteristics of Zone C in the Köppen system.

Köppen’s method is strongly influenced by the distribution of plant species (Griffiths, 1994). Criticism of Köppen’s classification has centered around the fact that it fails to address the level of incoming radiation and only indirectly allows for differences in the rate of evaporation (Lamb, 1972). Despite some of its shortcomings, Lamb (1972) points out

The classification has seen wide use and is likely known in nearly all countries. In the years after its introduction, modifications were made to the original Köppen classification to address some of its initial shortcomings.

In 1931, C. W. Thornthwaite developed a classification method based on climatic efficiency as related to plant communities (Thornthwaite, 1933). The classification is based on three parameters: (1) precipitation effectiveness (PE = 115sum[r_i/(T_i - 10)*(10/9)], where T_i = mean temperature (º F) of month i and r_i is its mean precipitation in inches), (2) temperature efficiency (TE = sum(T_i - 32)/4, where T_i = mean temperature (º F) of month i, and (3) the seasonal concentration of rainfall. PE is designated by a letter A – E, TE is designated by a letter A' - F', and the seasonal distribution of rainfall is broken down by: r for year-round precipitation, s for summer drought, w for winter drought, and d for year-round drought (Trewartha, 1954). The weakness of this system lies in the evaporation term, which completely ignores humidity and wind speed (Griffiths, 1994).

While there have been many climate classifications developed for the entire globe, there have also been many climate classifications developed for specific regions. In the 1950s, a climate zone map covering the 13 western states of the U.S. was developed (Sunset Publishing Corporation, 1995). This classification used winter minimum temperatures, summer high temperatures, length of growing season, humidity, and rainfall patterns as criteria for delineation.

The National Climatic Data Center (NCDC) produced climate classifications for the conterminous United States using twenty-four temperature and precipitation variables from each of the 344 climate divisions used by NCDC (Fovell and Fovell, 1993). The dataset consisted of monthly temperature means and precipitation accumulations for the conterminous United States over a 50-year period from 1931-1980. The NCDC classification delineated the 344 climate divisions into 8-, 14-, and 25-cluster groupings using principal components analysis and a hierarchical cluster analysis method known as group average linkage.

The United States Department of Agriculture (USDA) produced a classification scheme based on average annual minimum temperatures. This scheme resulted in a plant hardiness zone map for the United States, Mexico, and Canada that established boundaries in increments of 10º F (USDA, 1965). In the early 1990s, the classification was refined to establish boundaries by 5º F increments (USDA, 1990). The average annual minimum temperature from almost 8,000 stations for each of the years 1974 to 1986 in the United States and 1971 to 1984 in Mexico were analyzed. The map, computer generated by latitude and longitude, shows 11 different zones. Each zone represents an area of winter hardiness for agricultural plants.

DeGaetano and Shulman (1990) proposed a twenty-three zone climate classification scheme as an improvement to the USDA plant hardiness zones. As opposed to using one variable (winter minimum temperature), this scheme used maximum and minimum temperature, precipitation, wind speed, sunshine, relative humidity, growing season length, temperature extremes, growth units and elevation. PCA and cluster analysis, specifically the flexible clustering method, were used to identify areas of similar climatic conditions. The United States and Canada were divided into 1234 grid cells and the cluster solution was superimposed over the land mass. Since this classification integrated the effects of a wide range of meteorological variables, it may be better suited to classify plant hardiness.

In 1991, Douglas et al. (1992) completed a classification depicting agronomic zones for dryland winter wheat producing areas of Washington, Oregon, and Idaho. This classification used annual precipitation, soil depth, and growing-degree days from January 1 through May 31 as criteria to delineate the zones. Six zones were delineated and labeled (1) annual crop-wet-cold, (2) annual crop-wet-cool, (3) annual crop-fallow-transition, (4) annual crop-dry, (5) grain-fallow, and (6) irrigated. Each zone, regardless of its spatial location, had similar climatic and soil conditions therefore providing a common basis for the exchange of practices, technologies, and management decisions among producers, researchers, and extension personnel.

Two geo-climate zone classifications for the western region of the United States were produced in 1981 (Jallala, 1981). The first was based on eleven climatological variables: maximum, average, and minimum temperature of the coldest month of the year, maximum, average, and minimum temperature of the warmest month of the year, annual average temperature, number of frost free days, annual precipitation, January precipitation, and the July precipitation. The second was based on fourteen variables, which included the eleven climatological ones listed above plus longitude, latitude, and elevation. Meteorological data was obtained from the 1974 meteorological data books for each of the western states. The values of each variable for all the weather stations in each county were averaged resulting in 410 observations. Each observation represented a county, therefore, delineations followed county boundaries. The primary objective of the study was to identify geo-climate zones for the western region of the United States based on the similarity of natural elements. Twelve homogeneous geo-climatic zones were identified in order to provide for the efficient development of a coordinated research and/or extension program between agricultural experiment stations.

The task of developing agro-ecological zones for Africa was undertaken by the Food and Agricultural Organization (FAO) of the United Nations in the late-1970s (F.A.O., 1978). This classification applied to eleven crops distributed throughout the continent: wheat, paddy rice, maize, pearl millet, sorghum, soybean, cotton, phaseolus bean, white potato, sweet potato, and cassava. The zones were delineated based on growing period. Growing period was defined as the time during which temperature and precipitation were sufficiently available to permit crop growth (Griffiths, 1994).

In 1982, alfalfa-growing areas of the western United States were divided into fourteen climatically distinct zones by using cluster analysis (USDA, 1982). Data for latitude, longitude, elevation, frost-free days, mean temperature, mean maximum temperature, mean minimum temperature, and annual precipitation were collected from 30-year (1941-70) weather records. The 243 weather stations, located in or near alfalfa-growing areas were subject to cluster analysis resulting in a classification designed to benefit alfalfa workers in identifying climatological similarities and differences in growing areas of the western United States. The map accompanying this study did not attempt to delineate the entire western United States into areas with climatological similarities; it only classified alfalfa-growing areas of the western United States.

In 1983, seventy-six weather stations in New Brunswick were divided into eleven climatically similar clusters by means of multivariate statistical analysis on precipitation, various temperature parameters, elevation, latitude, and longitude (van Groenewoud, 1984). New Brunswick was then divided into climatic regions based on the clustering of the stations. The distribution of the seventy-six stations did not form a sufficiently dense pattern to be able to delineate the regions on their location alone. Therefore, climatic boundaries were drawn based on two rules: (1) in fairly level terrain the lines were drawn midway between the stations belonging to different regions and (2) in terrain with a steep elevation gradient the contour lines on a topographic map of New Brunswick were used to decide where the boundary lines should run. Comparisons of climatic maps with plant community distribution showed that climatic parameters might be a controlling factor in the distribution of vegetation. In a study using similar techniques, PCA and Ward’s clustering method were used to group sixty-three weather stations in Maine into four broad, homogeneous climatic regions (Briggs and Lemin, 1992). These four broad zones were then subdivided into nine sub-regions as part of a project to develop a productivity-oriented site classification system for spruce and fir. Due to the irregular and sparse distribution of the weather stations, the delineation of boundaries among climatic zones using the results from the cluster analysis alone proved to be difficult. As a result, a procedure involving stepwise regression was employed to determine the relationship between statistical scores for each weather station and the latitude, longitude, and elevation for each station. The climatic zones showed a high degree of correspondence with biophysical regions therefore reinforcing the multivariate analysis results.

Recent attempts to designate areas of climatic similarity are more concerned with customizing their studies for specific uses than their predecessors. Köppen’s climate classification was not greatly influenced by which sector of commerce (e.g., agriculture, forestry, etc.) would benefit most from the classification. Ideally, every industry would like to be able to derive useful information from a climate classification. The classifications by USDA, NCDC, DeGaetano and Shulman, etc. were of particular interest to industries such as agriculture, gardening, and forestry. The amount of information that is derivable from a climate classification for an industry is directly related to the goal of the classification (i.e., for whom it is tailored), the climatic variables that go into delineating the zones, and the way in which those variables are used. Ultimately, an ideal classification is specific enough to be beneficial to each individual industry while still being general enough to be useful to all industries.

The methods created over the years to delineate climate zones are many and varied. In general, to produce a climate classification map, weather stations or geographic areas are classified by hand or grouped statistically and then zonal boundaries are delineated manually, statistically, or in a GIS. Köppen arrived at threshold values after trial and error and improvement over a long period of years. The zones were strongly influenced by the distribution of plant species and the map was drawn by hand. Similar techniques were implemented by Thornthwaite in the early half of the century. In certain cases, threshold values were chosen and boundaries were superimposed over an area to display climatic extent (USDA, 1990; Douglas et al., 1992). A new crop of classification techniques began to evolve in the latter part of the century. Statistical analysis began to dominate climate classification methodology. In some instances, weather station locations were grouped and boundaries were extrapolated manually or by further statistical analysis (van Groenewoud, 1984; Briggs and Lemin, 1992). In other instances, geographic areas such as climatic divisions (Fovell and Fovell, 1993), counties (Jallala, 1981), and grid cells (DeGaetano and Shulman, 1990) were grouped statistically. Boundaries to delineate the zones followed geographic entities. Many of the recent classifications employ PCA and cluster analysis in some capacity (Fovell and Fovell, 1993; DeGaetano and Shulman, 1990; van Groenewoud, 1984; Briggs and Lemin, 1992). With the continued exponential increase in computer power and the integration of statistical software programs into spatial analysis systems, the latter techniques mentioned above will most likely be built upon to create classifications in the near future.