We have now completed DRB2070 Version 1.0, representing a baseline forecast of urban land cover in the Delaware River Basin out to the year 2070. A revised baseline and two alternative land cover forecasts are currently in progress and will soon be available. Click here to download the product description, or read more below.
Product overview:
DRB2070 Version 1.0 represents a baseline forecast of urban land cover in the Delaware River Basin (DRB) out to the year 2070. To develop these forecasts, we calibrated the SLEUTH urban growth model for the entire 43 county region of the DRB over the 2001-2006 time period, and validated the model for the 2006-2011 time period. We used the National Land Cover Database (NLCD) urban classes to represent urban land cover as developed or not developed (Figure 1).
Figure 1: All four NLCD developed classes were consolidated into a single representation to designate developed/non developed as required for SLEUTH input.
The primary input for SLEUTH is the exclusion/attraction layer, which describes areas that are more or less suitable for urban development. The exclusion/attraction layer developed for the DRB is the result of statistical and spatial modeling of accessibility, environmental suitability, employment and population spatial dynamics, and land protection (Figure 2).
DRB Modeling Subregions:
The Delaware River Basin is a large area, with the following characteristics:
- 43 overlapping counties in 5 states
- 35,000 sq. km (13,500 sq. miles)
- 8.2 million residents1
- 3.6 million payroll jobs2
- Provides water resources and ecosystem services to more than 15 million people, or 5% of the US population.3
Figure 2: The exclusion/attraction layer used to drive the DRB 2030 and 2070 baseline land use forecasts
We explored trends across all 43 counties in our study area to identify patterns in population, employment, and commuter flows, the main drivers of development on which we focus. Because of heterogeneous land cover dynamics across this large region, we subdivided the region into smaller, homogeneous modeling subregions to improve accuracy. We found that commuter flows between counties allowed us to group counties that share similar characteristics into eight different subregions. These subregions were named after the largest city within the region, where growth was focused, and modeled independently: Albany, Allentown, Baltimore-Annapolis, Delaware, Harrisburg, New York Metro, Philadelphia, and Upper-DRB (Figure 3).
Before running the model, the demand for new developed land by 2070 must be estimated. For the baseline land cover scenario, we calculated the average amount of development per person from 2001-2006 using Daily Human Intensity (DHI), or the sum of population and employment density. Finding the relationship between DHI and urban extent allows us to use population and employment projections to estimate the expected amount of development in 2070 for each subregion. For the Upper-DRB and Baltimore-Annapolis subregions, an inverse relationship was seen between urban growth and population growth; in other words, DHI decreased as urban land continued to increase. To model these regions, we linked growth with the region-wide DHI-urban relationship, relying on the assumption that factors other than employment and population affect growth in those regions.
Please note that the entire DelMarVa Peninsula was included in the Baltimore-Annapolis and Delaware modeling subregions in support of a separate project.
Figure 3: Subregions used to model land cover dynamics in the DRB and DelMarVa Peninsula
Baseline Land Cover Scenario:
The baseline land cover scenario represents recent trends in the Delaware River Basin for population growth, employment, regional build-out, regional infrastructure, and conservation efforts.
- Population Growth Trajectory: We based the population trajectory in the DRB on the EPA Integrated Climate and Land-Use Scenarios (ICLUS4) Basecase population forecast, which relies on moderate fertility, domestic migration, and net international migration rates, which reflect recent historical rates. Summary statistics for the resulting urban land cover change trajectory are presented for each modeling subregion in Table 1.
- Regional Build-Out Trajectory: our model considered accessibility to different resources: transportation (e.g. roads and intersections), urban density, and recreational resources (natural areas and water) as positive drivers to attract development.
- Regional Infrastructure Trajectory: We did not consider additional regional infrastructure. DRB2070 version 2.0 we plan to include current planned projects for road, rail, and energy infrastructure (electric and pipeline).
- Conservation Efforts: Non-forested wetlands are fully protected, forested or shrub wetlands have moderate to weak protection and we included protected lands as indicated in the PAD-US data.
- Sea Level Rise and Storm Surge Risk: Sea level rise and storm surge risk were not included in DRB2070 version 1.0. We will account for a global average of 6 feet (2 meters) sea level rise and Category 2 storm surge risk for the basin in DRB2070 version 2.0.
Urban land cover trajectory:
Using the NLCD, we calculated the total urban land cover in 2001 and 2011 for each modeling subregion, not just the portion of the DRB included in each subregion. Future urban land cover projections in 2030 and 2070 were obtained as the average of 100 Monte Carlo trials, and are summarized by subregion in Table 1. Download the DRB2070 version 1.0 product description for graphs of the urban land cover change trajectory calculated for each modeling subregion from 2001 – 2011 (observed from NLCD) and from 2011 – 2070 (DRB2070 Version 1.0 forecast) for the baseline land use scenario.
| Model subregion | Year | Developed land (ac) | Increase (ac) from observational period with standard error | Mean percent (%) increase from observational period |
| Albany | 2001
2011 |
48,343
48,802 |
–
459 |
–
0.95 |
| 2030
2070 |
49,412
51,230 |
610 ± 3.4
2,428 ± 13.5 |
1.25
4.98 |
|
| Allentown | 2001
2011 |
525,379
550,224 |
–
24,845 |
–
4.73 |
| 2030
2070 |
586,991
603,302 |
36,767 ± 39.8
53,078 ± 49.6 |
6.68
9.65 |
|
| Baltimore-Annapolis | 2001
2011 |
91,654
94,235 |
–
2,581 |
–
2.82 |
| 2030
2070 |
100,201
107,472 |
5,967 ± 15.1
13,238 ± 21.9 |
6.33
14.05 |
|
| Delaware | 2001
2011 |
211,508
230,360 |
–
18,852 |
–
8.91 |
| 2030
2070 |
291,254
330,970 |
60,894 ± 51.1
100,609 ± 72.0 |
26.43
43.67 |
|
| Harrisburg | 2001
2011 |
164,308
171,059 |
–
6,751 |
–
4.11 |
| 2030
2070 |
183,084
194,432 |
12,025 ± 20.6
23,372 ± 30.1 |
7.03
13.66 |
|
| New York Metro | 2001
2011 |
784,147
835,927 |
–
51,779 |
–
6.60 |
| 2030
2070 |
947,471
1,076,629 |
111,545 ± 561.6
240,702 ± 1,048.1 |
13.34
28.79 |
|
| Philadelphia | 2001
2011 |
1,133,757
1,191,562 |
–
57,805 |
–
5.10 |
| 2030
2070 |
1,343,191
1,607,660 |
151,629 ± 108.3
416,098 ± 250.3 |
12.73
34.92 |
|
| Upper-DRB | 2001
2011 |
98,614
99,488 |
–
873 |
–
0.89 |
| 2030
2070 |
100,913
103,528 |
1,425 ± 7.3
4,041 ± 15.7 |
1.43
4.06 |
Table 1: Developed acres and percent increase for the observational time period (2001-2011) and for forecasts of development in each modeling subregion in 2030 and 2070. For 2011, the increase and percentage is related to 2001. For 2030 and 2070, the increase and percentage are compared to 2011. Standard error is given in acres for the 95% confidence interval over 100 Monte Carlo trials.
DRB2070 version 2.0:
We are currently revising the baseline land cover forecast and preparing additional input layers related to energy infrastructure and climate change for DRB2070 version 2.0. Two alternative land cover scenarios will also be included in the next iteration.
1 US Census Bureau. American Community Survey Demographic and Housing Estimates, 2009-2013 American Community Survey 5-year Estimates.
2 US Census Bureau. Longitudinal Employer-Household Dynamics, 2013.
3 Delaware River Basin Commission. “Basin Information.” http://www.state.nj.us/drbc/basin/. Accessed March 9, 2017.
4 Integrated Climate and Land-Use Scenarios, version 1.3.2. County Population Projections. Environmental Protection Agency. https://www.epa.gov/iclus
