You are here

Harvard Forest Data Archive

HF290

Land-Cover Change Scenarios for Massachusetts 2010-2060

Related Publications

Data

Overview

  • Lead: Jonathan Thompson, Kathy Lambert
  • Investigators:
  • Contact: Jonathan Thompson
  • Start date: 2010
  • End date: 2060
  • Status: completed
  • Location: Massachusetts
  • Latitude: +41.3 to +42.9
  • Longitude: -73.5 to -69.9
  • Elevation: 0 to 1065 meter
  • Taxa:
  • Release date: 2017
  • Revisions:
  • EML file: knb-lter-hfr.290.2
  • DOI: digital object identifier
  • Related links:
  • Study type: modeling
  • Research topic: conservation and management; ecological informatics and modelling; regional studies
  • LTER core area: disturbance
  • Keywords: forest dynamics, future scenarios, land cover, land use, maps
  • Abstract:

    Working with a panel of practitioners and regional experts, we developed and analyzed four plausible but divergent land-use scenarios that depict the future of Massachusetts from 2010 to 2060. We simulated the land-use scenarios and their interactions with anticipated climate change by coupling statistical models of land use to the LANDIS-II landscape model and then evaluated the outcomes in terms of the magnitude and spatial distribution of (1) direct human uses of the landscape (residential and commercial development, agricultural, timber harvest), (2) ecosystem services (carbon storage, flood regulation, nutrient retention), and (3) habitat quality (forest tree species composition, interior forest habitat). Across all scenarios, conflicts occurred between dispersed residential development and the supply of ecosystem services and habitat quality. In all but the scenario that envisioned a significant agricultural expansion, forest growth resulted in net increases in aboveground carbon storage, despite the concomitant forest clearing and harvesting. One scenario, called Forests as Infrastructure, showed the potential for synergies between increased forest harvest volume through the sustainable practices that encouraged the maintenance of economically and ecologically important tree species, and carbon storage. This scenario also showed trade-offs between development density and water quantity and quality at the watershed scale. The process of integrated scenario analysis led to important insights for land managers and policymakers in a populated forested region where there are tensions among development, forest harvesting, and land conservation. More broadly, the results emphasize the need to consider the consequences of contrasting land-use regimes that result from the interactions between human decisions and spatially heterogeneous landscape dynamics.

  • Methods:

    Archived raster maps depict land cover for four scenarios at three points in time 2010, 2030, and 2060. For each scenario, we simulated landscape change within the entire 20,300 km2 of land area within the state of Massachusetts, U.S.A. (69.9– 73.58 E, 41.3–42.98 N) at a 0.25 ha grain size. While we tracked all land-cover classes throughout the state and the associated effects on various services, our focus was on the fate of the 12,800 km2 classified as forest in the 2005 land-cover map produced by and obtained from the state’s Office of Geographic and Information (MassGIS; http://www.mass.gov/mgis/). As a result, we simulated the area moving from the forest class to other land-cover classes, such as residential or agriculture, but did not simulate changes from non-forest classes in 2010 to other classes.

    We did not attempt to predict the precise location of future forest conversion or harvesting; instead we defined geographic probability zones based on social and biophysical parameters; the zones then dictated the total area affected by each of the land-use prescriptions (sensu Thompson et al. 2011). Within a probability zone, the spatial allocation of land-use prescriptions was random. The probability zones used in the Recent Trends scenario were developed based on regression tree analysis, which quantified the relationships between recent land-use change and a suite of potential predictors (e.g., distance from roads, population density, slope). For forest conversion to developed use, we analyzed all forest converted to developed uses in the period spanning 1999 to 2005 (following (DeNormandie 2009), which are the dates of the most current and accurate land-cover maps available. For timber harvest, we used data describing the location and intensity of all harvests in the state during the period spanning 2000 to 2010 (Kittredge and Thompson 2016).

    Land-use maps produced using the empirically-based regression tree of recent land-use trends provided a starting point for developing land-use maps for the alternative scenarios. The group subjectively manipulated the empirically developed regression tree from the Recent Trends to produce maps for each of the land uses in each of the remaining scenarios (Thompson et al. 2014, 2016). To do so, the group was guided through a process of adding new nodes to the tree, changing the suite of predictor variables or the values of nodes for existing predictor variables that, in turn, resulted in a map of the spatial configuration depicted in each of the storylines represented in the alternative scenarios (i.e. all but the recent trends scenario). We then produced new probability zone maps based on their decision trees (i.e. modified regression trees) and again allowed them to make modifications (Thompson et al. 2016). We iterated through this sequence until the participants were satisfied that each map represented the distribution of land uses throughout the state that they envisioned for each of their scenarios.

  • Use:

    This dataset is released to the public under Creative Commons license CC BY (Attribution). Please keep the designated contact person informed of any plans to use the dataset. Consultation or collaboration with the original investigators is strongly encouraged. Publications and data products that make use of the dataset must include proper acknowledgement.

  • Citation:

    Thompson J, Lambert K. 2017. Land-Cover Change Scenarios for Massachusetts 2010-2060. Harvard Forest Data Archive: HF290.

Detailed Metadata

hf290-01: MA scenarios land cover codes

  1. lucode: land use code
  2. lu.lc.desc: land use description

hf290-02: recent trends

  • Compression: zip
  • Format: img
  • Type: img file

hf290-03: opportunistic growth

  • Compression: zip
  • Format: img
  • Type: img file

hf290-04: forests as infrastructure

  • Compression: zip
  • Format: img
  • Type: img file

hf290-05: regional self-reliance

  • Compression: zip
  • Format: img
  • Type: img file