The landuse / landcover maps below represent a range of land use conditions,
from the historic (1850) to the present (1990), and beyond to five "Possible
Futures". In the Possible Futures, assumptions vary across the scenarios
in the categories of residential, agricultural,forestry, and reserve components
of the watershed. Lands in each of these categories vary by the amount of
land under one type of cover, the distribution of that cover, and the types
of practices on those lands. Click on the images below for map legends and
for more detailed discussion and visualizations of each scenario.
The watershed is comprised of three predominant land use types: low density rural residential, agricultural and forested lands. Consequently, these land uses figure largely in the set of assumptions guiding each possible future. These assumptions serve as the starting point for each mapped possible future. We address each below.
Human population demographers have long struggled with the challenge of producing accurate projections of population increases over thirty to fifty year planning horizons. To increase confidence in the plausibility of the possible futures envisioned in this project, the research team recommended that the stakeholders not simply accept a single population projection and use it in all futures, but instead use a range of projections each of which is consistent with the broad conservation or development-oriented nature of the scenario it appears in. To this end, residential lands in each future were influenced by a projected number of new people moving into the watershed, a number which varied from scenario to scenario. Represented by the number of new households added to 1990 levels, and calculated at 2.5 people per household, they range across the scenarios from a high of 1250 new households (above 1990 conditions) in the High Development Future, to a low of 125 new households (above 1990 conditions) in the High Conservation Future. These numbers are based on projections of two agencies, Benton County Development Department (Jim Hope, personal communication) and Portland State University's Center for Population Research and Census (PSUCPRC 1993) as well as the experience and judgements of the stakeholders.
Benton County Development Department projects 1000 new people (or 400 households) in the watershed by the year 2015. If projected to the year 2025 this translates into an increase above 1990 levels of 1500 people or 600 dwellings. Using Portland State University's Center for Population Research and Census projections for the growth rate of Benton County between 2015 and 2025, and assuming Muddy Creek watershed receives its area-weighted proportional share, this alters the projection to an increase of 1188 new people or 475 new dwellings by 2025 (PSUCPRC 1993). This became the basis for the residential portion of the Plan Trend Future.
The stakeholders used these figures and their own experience to set the full range of population projections, varying the number of new dwellings, as well as how they were distributed on the landscape. Among their decisions, stakeholders chose to concentrate new residential development in and around Alpine, currently zoned primarily for rural residential uses, where more social and physical infrastructure exists. This area came to be referred to as the `southern development zone' of the watershed. Its counterpart, a less compact area to the north, became known as the `northern development zone'.
Within the residential component of the possible futures, key variables are the total number of dwellings added relative to 1990 levels, the locations of those dwellings (compact versus dispersed), how their water supply needs are met, and whether or not all the new dwellings added are located within 1990 rural residential zones.
Agricultural land use across the scenarios vary primarily by the introduction of hybrid poplar as a short rotation paper pulp crop. The acreage devoted to hybrid poplar and the length of rotation (longer rotation poplar is grown for veneer) changes in each possible future. In the High and Moderate Conservation Futures, the introduction of windbreaks, hedgerows and stream-side buffers comprise an additional, significant variation in agricultural land use.
Forested lands are divided into public and private ownerships, with different policies governing each. Further, small privately owned forest lands are assumed to be guided by land management mandates different from the large, industrially managed private forest lands. Primary assumptions, which vary from future to future, governed the length of rotation (harvest) cycle and the size and distribution of harvest cuts.
Working with members of the research team, stakeholders decided to act as one group while at the meetings, instead of separating into their areas of expertise, such as farming or forestry. This was to help stakeholders learn about each others' concerns and facilitate the exchange of knowledge while at meetings. In response to this, research team members met individually with some stakeholders who were knowledgeable in specific areas, such as forestry or agriculture, to better reflect their experience and information in the different futures.
With the research team facilitating, stakeholders used paper maps, tracing paper and markers to develop locationally-specific depictions of land use and land cover pattern for each of the possible futures. Each future was reviewed and refined three times by the stakeholders and members of the research team, with each iteration producing a more complete mapped depiction. Because of busy schedules, attendance varied from meeting to meeting, however the group remained consistent throughout the process.
There are two general types of mathematical modeling approaches to depicting specific locations of future land use activities: deterministic and probabilistic (see Costanza and Sklar 1985). In deterministic modeling, a single set of inputs (expressed as map layers or land use siting rules) lead to a single conclusion. In probabilistic modeling, the inputs represent the likelihood of things occurring, but multiple runs of the same model will produce a range locations for the same land use reflecting the probabilistic nature of land use siting.
In examining the forested portions of the 1990 Land use / Land cover map of the watershed, it became clear that a simple deterministic description of industrial forestry in terms of average clear-cut size and frequency could not adequately describe the current state of the landscape. When deterministic algorithms were applied in describing the future landscape, the discrepancy became even more apparent: the landscape pattern which emerged was strikingly more regular than the stakeholder group thought reasonable or than has been produced by the past half-century of forest management in the watershed.
In the case of describing forest practices, the approach employed was to forecast future land cover patterns by specifying both standard practices and a "leakage" factor. The leakage factor represents the percentage of land owners not adopting standard assumed practices. We asked the stakeholder group to describe both standard practices and their level of adoption by different identifiable groups within the watershed (for example, the percentage of small private woodlot owners expected to clear-cut their timber holdings of a given age class in a given year). As employed, this leakage factor was higher, for example, for the small private forest land owners than for the large industrial private forest land owners in all scenarios. In our experience, this led to a more accurate characterization of the current landscape condition, as well as the possible futures.
The seven scenarios are available as a package here and individually from the detailed pages about each scenario. Clicking on the link button for the pdf file will initiate the download and launch the Acrobat reader if present on your machine. Clicking on the ERDAS button will cause a new panlel to be displayed here providing access to the individual maps and support files.
Adobe Portable
Document File (.pdf)
ERDAS-format GIS
Files (.gis) compressed using gzip.