The Development 2050 scenario assumes greater reliance on market oriented approaches to land and water use, with attendant emphasis on short-term economic gain in making land and water use decisions. It assumes recent trends in the relaxation of land use laws continue, resulting in fewer restrictions on where intensive land management may occur. Urban and rural residential uses expand significantly, with 1990 forest and agricultural lands making way for these expansions. The primary determinants of the land use / land cover patterns are commuting times from major employment centers, anticipated forest management mandates under more market-oriented assumptions, existing water rights, and projected crop patterns. Changes are projected in the amount, location, and pattern of urban, rural residential, agricultural, forest, and native vegetation land uses. Changes in water use are projected as the existing water rights associated with changing land uses are exercised. Federal reservoir management assumes the ca. 1990 operational pattern: reservoirs store as much water as possible in the winter and spring, release it in the summer to meet existing stream-flow targets along the mainstem of the Willamette River. The assumptions employed regarding each broad type of land and water use are described below
Willamette River Basin Population
Population projections for the PNWERC Alternative Futures Project for the Willamette River Basin (WRB) are based on Oregon Department of Administrative Services, Office of Economic Analysis population forecasts by county. These extend only to 2040, and PNWERC has extended the forecasts to 2050 using the regression line (rate of change.) The county forecasts have been adjusted to reflect the population within the area bounded by the Willamette River Basin based on 1990 population census blocks. The WRB's 1990 population is calculated to be 1,970,000 with 86% living within urban growth boundaries. The projected population for Plan Trend 2050, Conservation 2050 and Development 2050 is 3,900,000. In Plan Trend 2050, 93% of the population lives within urban growth boundaries; in Conservation 2050, 94% of the population lives within urban growth boundaries; and in Development 2050, 87% of the population lives within urban growth boundaries (Willamette River Basin Planning Atlas p.106).
A number of cities overlap the boundaries of two counties. County totals are combined for the purposes of calculating city and county projections where necessary.
Population forecasts for cities are based on Portland State University Center for Population Research data, supplemented by additional data from city and county governments ranging from 2020 to 2040. Where no outside sources exist, PNWERC extends the forecasts for the cities using individual city's rate of change, and matches county totals to within 1% of the DAS county total population. (details are available at: <http://www.fsl.orst.edu/pnwerc/wrb/metadata/devpop.xls>)
As with each of the three future scenarios, the projected population for Development 2050 is 3.9 million people, approximately double the 1990 population of the WRB. The Development scenario assumes 87% of these people will live inside urban growth boundaries, which have by 2050 expanded 129,000 acres beyond their 1990 extent. In this scenario, the future positions of UGBs are the result of a) a smaller percentage of people living inside urban growth boundaries (UGBs) interacting with b) larger lot sizes for new urban dwellings. Of the 573,000 total acres within Development 2050 UGBs, over 81% are developed as homes, stores, roads, and other built features, with less than 19% of the area inside 2050 UGBs vegetated. This dramatic UGB expansion over 60 years assumes new homes are built at densities somewhat higher than exists in 1990 (6.2 homes per acre basinwide for homes constructed 1990-2050 as compared to approximately 4.2 homes per acre basinwide existing in 1990), and by redeveloping and infilling only 5% of 1990 urban residential areas at this slightly higher density. In 1990, UGBs occupied approximately 6% of the WRB. In Development 2050 they occupy 7.8%, an average increase of over 2,100 acres basinwide per year for the 60-year period. The majority of this expansion occurs in the northern portion of the basin.
Populations in unincorporated and rural residential areas were derived from county data, and calculated based on available buildable lands. (details are available at: <http://www.fsl.orst.edu/pnwerc/wrb/metadata/devpop.xls>)
A key assumption of the Development Scenario concerns a significant increase in the total number of people living in rural areas. Consistent with this, the scenario assumes a general relaxing of restrictions on where new rural residences may occur. This is accomplished by excluding fewer areas from new rural development in this scenario than in the other two future alternatives. For example, new rural residences are not excluded from areas of Class I or II soils nor from riparian areas, nor from wetlands smaller than 5 acres. They are, however, excluded from the Federal Emergency Management Agency (FEMA) floodway (an area smaller than the FEMA 100 year floodplain), from industrial forestry parcels except when human population density exceeds 70 people per square mile, from parcels less than 1 acre in size and from parcels with less than 5 percent of their area or less than 1/4-acre suited for septic systems. The resulting pattern of 2050 rural residential land uses departs notably from trends experienced since the late 1970s.
In the Development scenario, agricultural land use declines in area relative to 1990 conditions due to dramatic increases in land area of urban and rural residential uses. While irrigated crop, grain, and hay uses decline basinwide in this scenario, other agricultural trends vary among the north, middle, and south basin counties. Grass seed uses increase in the north and middle basin, but drop by more than half in the south basin. Pasture uses increase in the north and south basin but drop by nearly one-third in the middle basin. Of the three scenarios, Development has the largest reduction in prime agricultural acreage with two-thirds of the converted lands occurring on prime farmland. The total area of land in agricultural production declines to 1.291 million acres in Development 2050, which equals approximately 17% of total basin area. Approximately 181,000 acres of 1990 agricultural lands are converted to other uses by 2050 (ref Berger and Bolte Evaluating the Impact of Policy Options on Agricultural Landscapes:An Alternative Futures Approach) Agricultural Land Use General information
A decision-making model was used to generate future representations of agricultural land cover. The model uses GIS-based spatial data sets as input to an object-oriented simulation model that computes dynamic agricultural field and basin scale attributes, and then performs a crop selection decision based on multi-attribute decision ranking. Decision constraints screened out unsuitable crops and then decision variables were used to rank the cropping system alternatives by applying the TOPSIS method (Hwang and Yoon, 1981), with the highest-ranking crop selected for the field.
This approach required the characterization of the initial agricultural system at the field-scale and the development of relative rankings of production level, profit margin, price variability, yield variability and management requirements for each of the cropping systems. The model has a yearly time-step, with special functions to integrate decadal updates from other components such as land use change or changes in water allocation.
Cropping Systems Selection:
The Willamette Valley supports a diverse selection of agricultural crops and management techniques. It is impossible to include all systems in our model, yet it is important to capture as much of the quality of this agricultural diversity as possible. To this end we modeled the following cropping systems for this study:
The rationale for this classification was to aggregate specific crops with similar production and management characteristics into crop classes and, with these crop classes, into rotation systems if applicable. Two cropping systems were further refined: Irrigated Annual Rotation had an early and late subclass reflecting planting date, while Irrigated Perennial had small and large acreage classes.
Creation of the Agricultural Fields Coverage: The agricultural landscape is made up of fields, considered here to be a parcel of land containing a single crop-type. The minimum area requirement for a field was 2 ha, which was used to differentiate commercially viable agricultural fields from hobby farms and gardens. Several data sources were used to define the boundaries of agricultural fields:
Manually digitized polygons
USGS Water-Resources Investigations Report 97-4268 associated data polygons
OWRD Irrigation Place of Use polygons
ODFW Land Use/Land Cover polygons
PNW-ERC Vegetation Classification, generalized to a one-acre minimum mapping unit (mmu.) and defined as discrete polygons.
Taxlot parcels for selected counties
Finally, each field was assigned area-weighted averages of mean monthly precipitation (Daly et al. 1994), available water capacity per foot of depth, (Natural Resources Conservation Service Soil Survey Geographic Database) and crop production potential ranking for each crop (Berger 2002). The crop production potential rank was determined by using yield data from NRCS county soil surveys in a supervised classification scheme that assigned a crop production class (very good, good, moderate, moderately low, low, and unsuitable) to each soil-crop combination.
Irrigation Requirements: Irrigation scheduling was based on monthly intervals, as described in the Western Oregon Irrigation Guides (Smesrud et al. 1997). This document also supplied the values for evapotranspiration, management allowable depletion, and effective root depth for each of the irrigated crops. On-farm irrigation efficiency was set to 80% for Irrigated Nursery Crops and 70% for all other irrigated crops.
Initial Distribution of the Agricultural Classes: The following land-cover maps were used along with crop and field data to define the initial (circa 1990) distribution of crops within the Willamette River Basin:
Oetter et al. 2001
Klock et al. 1998
Anderson et al. 1997
The assignment of a crop class to a particular field was based on both the confidence of the crop classification for a particular map and the total crop acreage for each county in the basin.
Irrigation availability - For a given field, irrigation availability was computed for each crop prior to crop selection. If at any time during the growing season there is forecasted to be inadequate irrigation, the crop was designated as unsuitable.
Crop Suitability - For a given field, a cropping system was deemed unsuitable if the crop production rank was unsuitable.
Market Demand - For the basin, an upper bound on crop acreage for each crop was set as the maximum acreage that occurred over the area during the period 1988-1998. This value was then increased or decreased yearly by the percentage change extrapolated by market trends. If the maximum acreage for a particular cropping system had not been fulfilled the cropping system was classified as suitable, otherwise it was classified as unsuitable.
Decision Variables: The attributes used to compare suitable crops represent the economic and management attributes of the cropping systems: crop production potential, price variability, yield variability, profit margin, and management requirements. Decision weights were assigned so that profit margin was considered the most important attribute, followed by management requirements, crop productivity potential, price variability, and yield variability.
Water availability within WABs was adjusted using WaterMaster (the computerized model of water allocation used in the PNW-ERC project) allocations.
For fields undergoing built development, field fragmentation was assessed using field shape and texture measures. The shape measures used were the perimeter-to-area ratio and fractal dimension of a field, while the texture measures were the angular second moment and entropy obtained from the gray level co-occurrence matrix (Haralick 1979) derived from the field. Any field with a shape value over four standard deviations beyond the mean value for the Circa 1990 agricultural fields was taken as an indication of conversion for the entire field. Cut-off values for texture were defined experimentally by simulating field fragmentation on sample fields and relating indicator levels to visual determinations of the suitability of the field for continued farming. Fields that had shapes or texture outside of conventional values were deemed to have been converted to a non-agricultural land use. These areas were assigned an increased probability for conversion to built uses or natural vegetation in subsequent iterations of other land allocation models.
An important assumption of the Development scenario regarding private forest lands is that, as population density reaches 70 persons per square mile, industrial forest parcels shift to become non-industrial forest parcels with attendant increases in rural residential land uses. There is also significant change in riparian vegetation policy, with 150-foot protection zones for large streams on federally managed forest lands, but no mandated riparian protection zones on state or private lands. Timber harvest rotation schedules are based on average annual cutting rates from 1973 to 1995 published in the May 1997 "Timber Harvesting Practices on Private Forest Land in Western Oregon," which equates to 60 years for private industrial and 128 years for non-industrial. The harvest schedule for State forest lands is based on a 100-yr. rotation, or 50% harvest probability. Modeled harvest units are 30 acres in size for federal, state, and private industrial lands, 5.6 acres in size for private non-industrial lands. Relative to Plan Trend and Conservation, industrial forests in the Development scenario place a greater emphasis on cutting of older timber and have reduced stream protection.
With forestlands continuing to occupy more than two-thirds of the basin, natural vegetation remains extensive under Development 2050 assumptions. While no explicit assumptions were stated in Development 2050 regarding non-forest natural vegetation, land use and management, primarily through land conversion to urban and rural residential uses, affect these areas of natural vegetation directly and indirectly.
Increases in the demand for surface water through 2050 reflect population and economic growth, as well as land conversion patterns, with per capita municipal demands assumed to be 12.5% greater in Development 2050 than in the Plan Trend 2050 scenario. Most increases in water demand occur within the constraints of existing ca. 1990 water rights and permits, with new permits available only for small self-supplied rural users and along the mainstems of the lower McKenzie and Willamette Rivers. The Development scenario for water evaluates the economic pressure on water resources. The supply of and demand for water are assessed in each of 178 "Water Availability Basins" (WABs), an administrative/hydrological unit defined by the Oregon Water Resources Department (OWRD).
The supply of surface water is represented in three sub-scenarios: a "normal" year, a "dry" year and a "wet" year. The supply is insensitive to changes in the landscape, so the three sub-scenarios of supply are the same across time, and across the alternative futures.
The demand for surface water is represented by existing surface water rights and permits. No new surface water rights are available in the Willamette River Basin in the summer, except along the lower McKenzie and the main stem of the Willamette River. Demand for surface water is summarized in six categories:
Under Development scenario assumptions, short-term economic gain has priority over long-term ecological function. Thus, working within the constraints of existing water law, land uses which employ water to produce greater short-term economic gains have priority in their demands for water over those land uses which employ water to produce lesser short-term economic gains. Allocation of water stored in federal flood control reservoirs is prioritized accordingly. Per capita municipal water consumption increases relative to 1990 conditions as urban industries with large water needs are added to the service districts of municipal water suppliers in major urban areas. Demands for water to meet agricultural needs change relative to 1990 conditions due to fewer 2050 acres in agricultural production due to urban conversion, and a larger percentage of 2050 ag acres being used for higher value crops with higher water needs. (Niemi et al., 2002)
Anderson, C.W., T. M. Wood, and J. L. Morace. 1997. Distribution of dissolved pesticides and other water quality constituents in small streams, and their relation to land use, in the Willamette River Basin, Oregon. Water-Resources Investigations Report 97-4268.
Berger, P. 2002. The generation and analysis of agricultural landscapes for futures research. Ph.D. Dissertation, Oregon State University.
Daly, C., R.P. Neilson, and D.L. Phillips. 1994. A statistical-topographical model for mapping climatological precipitation over mountainous terrain. Journal of Applied Meteorology, 33:140-158.
Haralick, R.M. 1979. Statistical and structural approaches to texture. Proceedings of the IEEE 67:786-804.
Hulse, D., S. Gregory, and J. Baker, editors, Willamette River Basin Planning Atlas, Oregon State University Press, Corvallis, Oregon, 2002.
Hwang, C.L., and K.P. Yoon. 1981. Multiple attribute decision making à methods and applications: a state-of-the-art survey. New York: Springer-Verlag.
Klock, C., S. Smith, T. O'Neil, R. Goggans, and C. Barrett. 1998. Willamette Valley land use/land cover map. Oregon Department of Fish and Wildlife, Informational Report. Available at: <http://www.nwhi.org>.
Niemi, E., D. Dole, and E. Whitelaw. 2002. Water Availability. Pages 114-16 in D. Hulse, S. Gregory, and J. Baker, editors. Willamette River Basin planning atlas: trajectories of environmental and ecological change. Oregon State University Press, Corvallis, Oregon, USA.
Oetter, D. R., W. B. Cohen, M. Berterretche, T. K. Maiersperger, and R. E. Kennedy. 2001. Land cover mapping in an agricultural setting using multi-seasonal Thematic Mapper data. Remote Sensing of Environment 76:139-155.
Smesrud,J.K., M. Hess, and J. Selker. 1997. Western Oregon irrigation guides. EM 8713, Oregon State University Extension Service, Corvallis, OR.