2020 Urban Growth Scenarios for Wetland Counties in the Central Valley of California

A Project for Ducks Unlimited

Bob Johnston
rajohnston@ucdavis.edu

Shengyi Gao
sgao@ucdavis.edu

Purpose

The purpose of this project was to project urban, suburban, and rural land development so that DU could take these data into their SVWM model, which is used to rate lands for acquisition for waterfowl use. Our urban growth projections would be viewed as a threat to such lands. Because DU believes that even one or two rural residential or agricultural activities could disrupt nesting and feeding sites, we had to attempt to project individual development projects throughout each county. Because of this exacting assignment, we acquired parcel data for almost all counties of concern. We also took care to use the Calif. DWR land cover data, which identifies many rural farmsteads. The parcel data took many days per county to classify, as every county uses different data categories. We added a Residential Very Low Density to UPlan for this project, so that we could use Resid. Low to represent large-lot parcels near to cities and towns. We allocated RL with our attraction weights. Then, we used Resid. Very Low for true rural residences and farms and allocated them randomly. We took care to use the Census block population data for 1990-2000, overlaid onto parcel boundaries, to get the percent of county population growth in RL and RVL, in order to calibrate these numbers of dwellings in each density category.

Aggregation of land use types in general plan

The names of land use designations and residential densities are generally different in different cities and counties. The reclassification and aggregation of land use designations in general plans are generally necessary. In this project, all industrial land use designations in general plans are designated as industrial. High density commercial includes commercial buildings with three stories or more (downtown commercial, highway commercial, regional commercial center, commercial park, etc.) Low density commercial includes community commercial, neighborhood commercial, general service commercial, etc. The densities for residential are defined as follows:

• High density residential: more than 8 dwelling units per acre
• Medium density residential: 0.5 - 8 dwelling units per acre
• Low density residential: the lot size is between 2 and 10 acres.
• Very low density residential: the lot size is more than 10 acres.

Identification of LDR and VLDR

The LDR and VLDR in UPlan usually are not differentiated and both are designated as agriculture in county general plans. They are designated separately in zoning maps, but, unfortunately, digital zoning maps are not generally available. Therefore, parcel maps becomes the only alternative to obtain LDR and VLDR designations. Generally speaking, as the population goes up, the parcels will be split. The parcels whose current area is less than 20 acres will be probably split into more than two smaller parcels and become low density residential areas. The parcels whose current area is between 20 acres and 40 acres may be very low density residential if they are split one time or become low density residential if they are split two or more times. "Hot spots," i.e. the parcels that are located in the census blocks with positive growth of population 1990-2000, and the parcels closer to urban areas, have a higher probability to be split. To modify the digital general plans to differentiate LDR and VLDR, we cut the "agricultural" from the general plan to form a new layer, and then use the new layer to clip the parcel layer. The layer clipped from the countywide parcel layer is reclassified according to the zoning, if available, and the parcel size. The reclassified layer will be combined with the general plan to form a new general plan with LDR and VLDR designations

Identification of ratios of residential types

To get the ratio of HDR, MDR, and other (LDR and VLDR, together) residential land use types, we used the housing data of the California Building Industry Association (CBIA) and the housing stock data of the State DOF. As to the ratio of low density residential and very low density residential, we use the net growth ratio of population between 1990 and 2000 in LDR and VLDR. First, we use the general plan, DWR urban, and farmland developed as references to find the blocks that represent the urban blocks (including all land uses, except LDR and VLDR) on the census block map (population change, 1990-2000), and then delete them from the map. Second, we identify the rural communities with medium density residential land use on the parcel map and use them as references to find the matching census blocks, and delete those blocks from the census block map. Third, we use the LDR and VLDR obtained in step 2.2, below, to identify the census blocks with LDR and VLDR growth, 1990-2000. Taking the absolute growth of persons in LDR blocks and in VLDR blocks, we get the growth ratios of LDR/total residential and VLDR/total residential between 1990 and 2000. They will be used as the UPlan inputs for proportions of residential land use types for 2000-2020.

Rules of allocation

In the UPlan model, the users have the flexibility to define their own rules to allocate the new growth within the general plan to test the impact of land use policies. In this project, through the aggregation, the general plans are "standardized" in terms of the same land use categories to all counties. We also apply the same rules to allocate the growth. The allocation of new growth into general plan categories is as follows:

Attractions and masks

In this project, freeway ramps, highways, major arterials, minor arterials, city sphere of influence (SOI), and the census blocks that have net growth of population in low density and very low density areas (determined from parcel sizes) are used as attractions. An SOI is where infrastructure will be provided within the next ten or fifteen years, and so is given a higher weight than ramps or roads. The number of buffers of each attraction layer and the weights of each buffer are listed in Table 1.

The masks include streams, lakes, public lands, existing urban, wetland reserve program (WRP) areas, and deeded permanent easement areas (only in Merced County). The existing urban layer is a combination of DWR urban and parcel urban. DWR layers are very helpful, especially when we have no other data sources, to identify rural low density residential and very low density residential. Conceptually, the parcel layer is the best source for existing urban. Ideally, a parcel layer will have information on general plan land use, existing land use, zoning, potential parcels, land values, improvement values, structure area and value, and will be updated every month, or at least twice a year. On such an ideal parcel map, it is very easy to get good existing urban and urban vacant land layers. Unfortunately, not all counties have parcel data, and not all counties are willing to share their parcel data without charge. The parcel data we have are far from ideal. It takes a long time to interpret the parcel data to get existing urban.

Interactive GIS Maps:

This project is funded by Ducks Unlimited. For more information, please visit: http://www.ducks.org/