- Data & Archives
- Policy & Conservation
- Education & Opportunities
- News & Events
You are here
Harvard Forest Data Archive
Land-Use Impacts on Ecosystem Services Provisioning in Massachusetts 2001-2011Related Publications
- hf245-01: land use impacts
- Lead: Jonathan Thompson, Meghan Blumstein
- Contact: Information Manager
- Start date: 2001
- End date: 2011
- Status: completed
- Location: Massachusetts
- Latitude: +42.0 to +43.0
- Longitude: -73.0 to -70.0
- Elevation: 0 to 1062 meter
- Release date: 2015
- EML file: knb-lter-hfr.245.4
- DOI: digital object identifier
- EDI: data package
- DataONE: data package
- Related links:
- Study type: modeling
- Research topic: conservation and management; ecological informatics and modelling; regional studies
- LTER core area: disturbance
- Keywords: biodiversity, conservation, ecosystems, future scenarios, land cover, land use, modeling
Meeting fundamental human needs while also maintaining ecosystem function and services is the central challenge of sustainability science. In the densely populated state of Massachusetts, USA, abundant forests and other natural land cover convey a range of ecosystem services. However, after more than a century of reforestation following an agrarian past, Massachusetts is again losing forests, this time to housing and commercial development.
We used land-cover maps, ecosystem process models, and land-use data bases to map changes (2001, 2006, 2011) in eight ecosystem service variables and to identify “hotspots,” or areas that produce a high value of five or more services, at three policy-relevant spatial scales. Water-related services (clean water provisioning and flood regulation) experienced local declines in response to shifting land uses, but changed little when measured at the state-level. General habitat quality for terrestrial species declined state-wide during the study period as a consequence of forest loss. In contrast, climate regulation (carbon storage) and cultural services (outdoor recreation) increased, driven by continued forest biomass accrual and land protection, respectively. Timber harvest volume had high inter-annual variability, but no temporal trend. The scale at which hotspots are delineated greatly affects their quantity and spatial configuration, with a higher density in eastern Massachusetts and 10–12% more hotspots overall when they are identified at a town scale as compared to a watershed or state scale.
Ecosystem service hotspots cover a small percentage of land area in Massachusetts (2.5–3.5% of the state), but are becoming more abundant as urbanization concentrates ecosystem service provisioning onto a diminished natural land base. This suggests that while ecosystem service hotspots are valuable targets for conservation, more are not necessarily better since hotspot proliferation can reflect the bifurcation of the landscape into service and non-service provisioning areas and subsequent loss of diversity across the landscape.
This section describes the data and R and Python scripts needed to reproduce our findings. References are to directory and file names in the Zip file (hf245-01-land-use.zip).
Land-Use Land-Cover Maps
The initial LULC maps (Initial_LULC_Files) were originally sourced from the NLCD website (https://www.mrlc.gov/), then transformed into a common projection and extent, and re-classed using Code/Step_1_Change_LULC_Classifications.R.
Water Services Analyses
Water service analyses were performed using the InVEST (3.0.1) standalone models (https://naturalcapitalproject.stanford.edu/software/invest). All initial data files are included in the Water_Services_Inputs directory. Step 1: Run Stand-alone InVEST module (water yield, water purification). Use Initial Data Files in Water_Services_Inputs: biophysical_table.csv, Eto_All.img, MA_precip_all.img, soil_depth.tif, soil_pawc.tif, watersheds.shp, subwatersheds.shp, water_purification_threshold.csv. Step 2: Modify nitrogen and phosphorus outputs for use in the hotspot analysis using Code/Quantifying_Services/ Step_2_NP_Create_Retention_div_Export.py.
Habitat Quality: Habitat quality analyses were performed using the InVEST (3.0.1) standalone model. Initial files can be found in Habitat_Quality_Inputs. Step 1: Create Threat Rasters. Use Initial_LULC_Files as initial input and Code/Quantifying_Services/InVEST/Habitat_Quality_Step_1_Threat_Rasters.R. Step 2: Threat rasters and initial files (Habitat_Quality_Inputs) are run through InVEST. Access_01.shp, Access_06.shp, Access_11.shp are layers derived from the open lands source file on Mass GIS (https://www.mass.gov/orgs/massgis-bureau-of-geographic-information). Also use sensitivity.csv, threats.csv. Step 3: Modify Degradation Outputs to Habitat Quality. Code at Code/Quantifying_Services/InVEST/Habitat_Quality_Step_3_Inverse_HQ.R.
Aboveground biomass quantified using Landis-II (http://www.landis-ii.org/) and the biomass succession module using methods and parameters described in Thompson et al 2011 Ecological Applications 21: 2425-2444. Step 1: All Landis input files needed to reproduce aboveground biomass projections are on Carbon_Initial_Files. Step 2: Adjust biomass outputs from Landis using code at: Code/Quantifying_Services. Run Step_2_Carbon_Postprocessing_Routine_1.py using Landis outputs. Step 3: Adjust Landis outputs with the land cover specific values using code at: Code/Quantifying_Services. Run Step_3_Carbon_Postprocessing_Routine_2.R.
The Massachusetts Department of Conservation and Recreation maintains the state timber harvest database and shared the data with the authors according to a memorandum of understanding with Harvard University, which stipulates that the data not be shared. Please contact the authors or Jennifer Fish at MA DCR (413-545-5753) if you have questions. Step 1: Process initial data and match to tax assesor data using Timber_Harvest_Initial_Files/Large_Tax_Parcels_wtwns.shp acquired from MASS GIS and assembled using Combine_Parcel_Shapfiles.py at Code/Quantifying_Services. Step_1_Harvest_Process.r cleans the data and matches the addresses from the tax assesor parcels to the timber database addresses by road. Step 2: Randomly select parcels for harvest by street using code at: Code/Quantifying_Services/Step_2_Harvest_Select_Parcels_by_LOC.R.
Recreation files were all downloaded from MassGIS and adjusted by hand in ArcGIS to get them all in the same projection and extent. Step 1: Reclass the individual recreation files: using Code/Quantifying_Services/Step_1_Reclass_Recreation_Data.R. Step 2: Combine all the rasters and process for the final recreation outputs using Code/Quantifying_Services//Step_2_Processing_Recreation.R.
Initial Impervious data downloaded from MassGIS and mosaiced in ArcGIS (https://www.mass.gov/orgs/massgis-bureau-of-geographic-information). Step 1: Create impervious rasters. Use Initial_LULC_Files/Water_Services_Inputs/Impervious_2005.img and code at: Code_Quantifying_Services/Step_1_Calculate_Impervious_by_LULC.R.
All hotspots code can be found in Code/Hotspots. The folder also contains Expand_Clip_extent.py, which can be a useful tool in making sure the data are all the same size and extent (which is necessary for the analyses to work). Step 1: Identify all high-value areas for each ecosystem services (top 20th percentile) using Step_1_Hotspot_Identification.R. Step 2: Sum all of the high-value rasters to get the hotspot output rasters using Step_2_Sum_For_Hotspots.R. Step 3: Quantify hotspots and create tables of HA of land in each class (0-8) using Step_3_Quantify_Hotspots.R.
All of the code that was used to make the figures is included in Code/Figures.
This dataset is released to the public under Creative Commons CC0 1.0 (No Rights Reserved). Please keep the dataset creators informed of any plans to use the dataset. Consultation with the original investigators is strongly encouraged. Publications and data products that make use of the dataset should include proper acknowledgement.
Thompson J, Blumstein M. 2015. Land-Use Impacts on Ecosystem Services Provisioning in Massachusetts 2001-2011. Harvard Forest Data Archive: HF245 (v.4). Environmental Data Initiative: https://doi.org/10.6073/pasta/f0a1d3f3c139666ac281d26f2383682e.
hf245-01: land use impacts
- Compression: zip
- Format: Esri shapefile, text, Python script, R script
- Type: document, script, vector GIS