Inferring the Local Social Cost of Wind Turbines

This project estimates the valuation of spatial attributes as implied by the Lower Austrian wind power zoning. Data processing uses the cleo package.

Data

The analysis includes proprietary data which must be obtained from:

• Important Bird Areas: BirdLife International via BirdLife Data Zone
• Overnight Stays in Austrian Municipalities: Statistics Austria (request via [email protected])

All other data are downloaded from sources referenced in:

from scow.site_data import generate_data_dict

For the World Database on Protected Areas (WDPA), if the provided link is broken, find the latest link at Protected Planet.

Installation

The code is developed using python 3.9 and relies heavily on the cleo package. Other dependencies are listed in environment.yml. To create a Python environment named myenv with these dependencies:

conda env create -n myenv -f environment.yml

For cleo installation and usage, refer to the cleo GitHub page.

Apart from python, installations of R and GAMS are also required.

Execution

Configuration

Configure the following settings in config.py:

repo = Path("path/to/repository")
data_ver = "2014"  # "2014" for the time of zoning decision or "modern" for current wind turbines

num_runs = "2500"  # number of iterations for discrete choice estimation
selection_criterion = "BIC"  # options: "AIC" or "BIC"

spacing = 3  # minimum pixels between wind turbines for optimal location
nturbines = "auto"  # automatic selection of deployed wind turbines

rdir = Path("C:/myprogs/R/R-4.2.2/bin/Rscript")  #  path to Rscript.exe (Windows)
gamsdir = Path("c:/myprogs/GAMS/45")  # path to folder containing gams.exe (Windows)

Data Processing

Execute the scripts sequentially:

1. 01_spatial_data.py: Retrieves and preprocesses wind resources data from the Global Wind Atlas and other spatial characteristics, saving dc_data_2014.csv and dc_data_modern.csv in data/processed.

2. 02_estimate_wtp.py: A Python wrapper around R code in discrete_choice.R, producing files like spatialdc_coefs_{datafile}_{run_name}.csv in data/results.

3. 03_postprocess_wtp.py: Computes social costs, plots maps of private and social costs, and the distribution of various costs.

4. 04_optimal_siting.py: Solves the location optimization problem, saving results to opt_locations_{data_ver}_{objective}.csv and opt_cost_{data_ver}_{objective}.csv in data/results.

5. 05_postprocess_sites.py: Processes optimal wind turbine sites, computes local and total social costs, and plots maps of optimal locations and the social cost curve.

Logging

Logs are kept in data/logfile.log.