This document describes how the grids modeling the Willamette River Basin "Plan Trend (01/15/00, Version 6)" were constructed. These scenarios built upon a base condition described by the grid "Land Use/Land Cover ca. 1990 (Version 12/15/99)."

The approach taken was to develop separate grid layers for each time step which were then merged together in a specific order to produce the final grid representing land use/land cover for that time step.  Viewed as a vertical stack, the order of assembly of a composite grid for a time step is displayed in the table below: the uppermost grid components retain their values, blotting out spatially coincident values of lower grids.  This implements the precedence of themes.

This process follows closely that used to develop the base condition of 1990.  For each decade from 2000 to 2050, land use/land cover changes for forests, agriculture, riparian buffers, and urban and rural expansion were modeled resulting in a thematic grid for each land use that contained cumulative changes since 1990.  These were "sandwiched"  with static components of land use (the same grids that were used in the construction of the 1990 base condition), yielding a composite grid of land use/land cover for the decades 2000 through 2050.

Each of the thematic grid layers are referenced to the Basin DEM,  and are valued according to the LULC legend. The background/undefined cells are set to NODATA.  The composite grid is similarly referenced to the Basin DEM, and is clipped to the Basin boundary.
 
 

PLAN TREND

CATEGORY /Source	Legend Value	Source and Commentary	Grid name	Construction Script name
Major Roads  Source:  dot97c                 Static	19-21	[dot97c, igds-level 50 => legend 19, igds-level 51 => 20, igds-levels 52, 53 => 21]  Level 50 roads are 3 cells wide; level 51 are 2 cells wide; levels 52-53 are 1 cell wide. (Level 50 represents interstate highways; 51 represents principal arterials; 52, minor arterials; 53, major/urban collectors)	roads19-21	--
Minor local roads within ugbs in counties other than Yamhill and Columbia Source: Taxassessor-collector parcel data         Static	21	[taxassessor-collector parcel data]  Roads are described as spaces between taxlot polygons within the tax-assessor-collector data sets; these polygons are gridded with the number of cells representing the roads dependent on the polygonal description.  This description is not available for cities in Yamhill and Columbia Counties because of the absence of digital tax-parcel data.	ugbtaxrds	--
Minor local rds within all ugbs intersecting Yamhill and Columbia Cos.  Source: dot97c                       Static	21	[dot97c igds-level 12 (local roads) clipped to ugbs]  This layer is 1 cell wide. In the absence ov taxlot data for Yamhill and Columbia, an alternate means is needed to represent the local roads within their ugbs. Comparison of dot97c igds-level 12 with the Eugene-Springfield taxlot data shows a good alignment with the taxlot right-of-ways that are roads. Thus, by using these ODOT data, we are confident that the treatment of roads in Yamhill and Columbia Co. ugbs will be consistent with those ugbs for which we do have taxlots and from which we obtain roads directly.	dot21ugbyc	--
Railroads Source: dot97c                         Static	18	[dot97c, igds-level 18 (railroads) edited to remove the cross-hatching arcs]  This is gridded 1 cell wide.  When railroads and roads run side-by-side, and in close proximity, the gridded representations often overlay.  The railroad layer was selected to lie "beneath" the roads grids so that the roads representation would be continuous.  This means that the gridded railroad lines may "disappear" beneath the gridded roads in places.	rr18	--
Channel, non-vegetated: inside mainstem of Willamette River Source: TM landcover data/FSL and Willamette River active channel 1995                             Static	29	[legend 26-28 (built) inside the ac1995p feature boundaries(see water below)].  This was extracted from the TM landcover so that "built" features within the Willamette River would be represented.  Islands within the Willamette River (as defined by OSU coverage ac1995p) lie within this grid depiction, so that any "built" categories on those islands are given this legend category (29).  "Built" landcover also occurs within the water of this river coverage.  It is assumed that most of the occurrences in this category are gravel or channel bars (according to Doug Oetter  (pers comm April 1999)), with some revetments, river-side roads, and, in the METRO area, bridges and piers.	chan29	chan29.aml
Water Source: TM landcover data/FSL, river network and Willamette River active channel 1995                                                                       Static	32-33	[TM landcover water (legend 33) + ac1995p code <> 11 (legend 32) + rivreach_2 Strahler order >= 5 (legend 32).] The TM landcover water was found not to represent rivers and streams adequately; the goal was to have major rivers that were likely to be wider than 15m continuously defined. "AC1995P" (polygon coverage from the River Restoration Project) was used so that the width of the Willamette mainstem would be described adequately.  (Islands were excluded from this representation and thus, the veg and other categories layered are depicted on these features).  "Rivreach_2 " Strahler orders 5 and greater were gridded at 1 cell.  Note that there are some inconsistencies between coverage ac1995p and rivreach_2 in the description of the mainstem.  Thus, in some areas, a river loop (from rivreach_2) will add to the water area defined by ac1995p. In most places, however, the channel location defined by rivreach_2 is enclosed within ac1995p. The order invoked in creating this grid was TM landcover water (legend 33) over "AC1995p" (legend 32) over "rivreach_2" Strahler 5-7.	wat3233	wat3233.aml
Rural structures, outside ugbs                 Modeled	16	[digitized from usgs quads, and modeled] These structures represent rural dwellings and rural structures. Structures derived from scanned images of USGS quads minus those that fall beneath the gridded representation of roads, railroads and water features represent the 1990 condition. These persist through time. Development in rural residential zones (only) adds dwellings over time.	.	.
Turf grass, inside ugbs Source: TM landcover data/FSL                                                                                                   Static	91	[turf grass but primarily golf courses, from TM land cover category 91.] This was separately extracted from the landcover so that it would overlay the urban landuse data obtained from taxlot information (see below). This was done because it was found through comparison of the taxlot information and the "parks" (landcover category 91) that almost all of the occurrences of "91" were of golf fairways. The civic/os category (above) will include public golfcourses but exclude private courses. It was decided that fairways would be considered unbuildable (not vacant) in the urban modelling, and that they should not be described as "commercial" (a common landuse designation given in the tax lot data to private country clubs, etc.) -- such a commercial designation with no employees would bias the commercial lands inventory used in the population distribution.  Also, by explicitly designating fairways, this vegetative cover would be available for consideration as habitat within ugbs.  By placing this layer above the taxlot data, private fairways will be identifiable. Outside the ugbs, the Turf Grass category will be present in existing conditions (1990) only as it is not masked out by any agricultural layers.  Ugb boundaries may expand to encompass turf grass areas outside of ugbs in the futures. However, this included turf grass is not given a sacrosanct standing-- it is assumed that turf grass outside ugbs in 1990 is primarly an agricultural crop and is only protected from development if it falls within a civic/os area.	turf91u	--
Civic/open space Source for boundaries: multiple -- taxlots, parks, and ownership databases                                                                                                                   Static boundaries/ Modeled land cover	11, 24 and >=29	[Boundary datafrom taxlots, METRO/RLIS parks, ownerd;Landcover data from static and all modeled layers listed below excluding urban] This grid contains the landcover for all parcels defined as civic open space within the WRB, areas that are off-limits to urban and rural development.  Such areas are expected to be potential natural reservations within the WRB.  These data include the following categories: public schools, buildings, colleges/universities, Federal, state, county and local parks, wildlife refuges, reservations, public and private cemeteries, nature conservancy reserves. The federal and state forests are EXCLUDED from this category as these forests are considered to be primarily focused on wood production.  Inside the ugbs, the "built" categories are mapped as "urban built unknown" (legend 11); outside the ugbs, the "built" categories are mapped as "rural built unknown" (legend 24); the rest of the landcover within the civic/os area remains as defined by the amalgamation of the other landcover data (as layered below):the "natural veg", the ag initial conditions (see below), the amended TM data, and the TM landcover data. This layer will have areas that coincide with taxlot defined areas (see "urban lu" below): this grid overrides those definitions. For example, in the taxlot definition, schools are defined as commercial; if these are public schools, civic/os redefines them to whatever the amalgamated landcover depicts. Thus, a school with a building and lots of playing fields will show up as partly "urban non-vegetated unknown" (11) and partly grass/forest etc.  The land cover within the civic/os boundaries does not remain static: changes in conifer age, ag crops, and riparian buffers can occur within civic/os boundaries.	civic12	civic12.aml
Urban landuse inside ugbs                   Modeled	1-4, 6, 7, 8, 10	[lowdensity 0-4 du/ac = 1, medium low density 4-9 du/ac = 2, med high 9-16 du/ac = 3, high >16 du/acre = 4; 6=commercial, 7=comm/ind, 8=industrial, 10 = mixed comm/residential] Urban landuses as modeled based on population growth, siting preferences, and zoning restrictions. Urban growth boundaries are expanded according to stakeholder advice.	.	.
Remapped built TM data Source: TM landcover data/FSL and  household density  derived fromU.S. census                                                                                             Static	1-4, 7, 11, 24	[landuse derived from remapping TM landcover categories 26-28 inside ugbs and outside ugbs; census data]. Inside ugbs: built high density landcover was remapped to commercial/industrial = 7; built low and med densities were remapped to residential, and the Census household density grid was then applied to determine the residential density applicable to that residential grid cell. This technique was used primarily for those counties where taxlot data were not available (Yamhill and Columbia Cos.). This remapping is also used to represent the landcover built categories that fall within undefined or ambiguous parcels in ugbs with taxlot data (as described above). For parcels defined as "vacant" by the taxlot data, any "built" landcover is classified as "urban non-vegetated unknown" (11); similarly, for civic/os areas.  Outside ugbs: "built" landcover is mapped as "rural non-vegetated unknown" (24). According to Doug Oetter (pers. comm, April 1999), the landcover built density categories may include fallow land, rock, bare ground, gravel pits, gravel bars, roofs, roads, etc. Here, it is assumed that the rural structure data, the ODOT roads, etc. (as described above) are better descriptors of the land uses we wish to isolate. Everything that is then left over after these land uses blot out the landcover "built" is set to "rural non-vegetated unknown". Note that gravel pits and mines will most likely be in this category (except for any that fall inside the mainstem Willamette River coverage, as described above).	bltremap	bltremap.aml
Conifer forests   Modeled	56-61	Forest conifer cover classes as modeled based on owner, age, and cutting regime	.	.
Riparian buffers Modeled	62, 80	Woody or grass buffers added to edges of agricultural fields along 303-d listed streams	.	.
Conifer/deciduous/mixed forests,  flooded marsh, natural shrub Source:  TM landcover data/FSL       Static	51-62, 89, 87	[data from TM landcover data/FSL] These data represent vegetation in the amended+disturbed+TM landscape that is retained in place of spatially coincident ag cover (below). All forest classes, natural shrub, and flooded marsh pixels are selected. "Natural grass" is not retained as it's classification is of low confidence.	vegaboveag	vegaboveag.aml
Amended ag crops in proximity of urban and rural development      Modeled	85	When urban and rural development begin to dissect an agricultural field and a critical threshold of conversion from agricultural use is exceeded, the land cover of that field is converted to pasture.	agamend	.
Agricultural lands Source:  Bolte and Berger, Bioresource Engineering, OSU   Modeled	66-68, 71-73, 83, 85, 90, 92, 93, 95	[from John Bolte and Pat Berger, OSU] Agricultural cover crops as described by Bolte&Berger for the modeled decade (Version 3a, dated 12/6//99).	ptag<decade>	--
Amended TM landcover         Static	49	Replacement of landcover data defined by TM analysis (refined by forest disturbance layer, see below) due to low confidence in original classification: orchards within ugbsare reclassified here into the more general "urban tree overstory"	amendtm	amendtm.aml
Forest Disturbance Source: Forest disturbance/FSL, Cohen     Static	56	[data from Cohen  et al, FSL] Areas of forest disturbance from 1972 to 1988 that introduce conifers of age 0-20 years; water, "turf" inside ugbs, and any built categories from TM data have precedence over the disturbed forest category.	disturb72-88	--
Landcover TM Source:  TM landcover data/FSL     Static	26-28, 33, 39-40, 42, 51-62, 73-93	[data from TM landcover data/FSL] Remote-sensed landcover from TM data/FSL with classes renumbered to match the master LULC legend. This is the 40-class version which includes conifer forest age classes.	lc90_4	--