A Dataset Cataloging Product-Specific Human Appropriation of Net Primary Production (HANPP) in US Counties
This dataset is associated with the paper “Product-Specific Human Appropriation of Net Primary Production (HANPP) in US Counties” (Paudel et al, 2023). This dataset comprises Human Appropriation of Net Primary Productivity (HANPP) values for 3101 counties in the conterminous US for the years 1997, 2002, 2007, and 2012. For this dataset, HANPP is the carbon content of specific crop, timber, and livestock grazing products appropriated by humans in a county in a year. To calculate HANPP, raw agricultural data were downloaded from public databases such as MODIS, Landsat, Cropland Data Layer, USDA National Agricultural Statistics Service, and US Forest Service. Raw data were processed in Microsoft Excel using stoichiometry derived from established scientific literature. For more information about the sources and methods used to produce this dataset, please see the accompanying Data in Brief paper "A Dataset Cataloging Product-Specific Human Appropriation of Net Primary Production (HANPP) in US Counties" (Paudel et al 2023).
Steps to reproduce
To calculate total HANPP for each county, we calculated three subsets of HANPP (crops, timber, and grazing) and added them together for each county for the years 1997, 2002, 2007, and 2012. Raw crop data are from the U.S. Department of Agriculture. For counties without available crop data we used similar crop data to estimate. We converted crop yield estimates to HANPP. We started with a formula for calculating the NPP harvested during crop production from data on economic yields. We identified the ten largest crops and all other crops aggregated together. We assumed that the onsite HANPP density of minor crops is the national mean for major crops and is constant. We used stoichiometry formulas to calculate the NPP harvested for each crop in each county. There are two units of HANPP: county-wide and onsite. HANPP was further divided into used and unused portions and above and belowground portions. Raw data on softwood and hardwood harvests for each county were obtained from U.S. Forest Service data. We converted timber harvest measurements from thousand cubic feet to HANPP. We adapted the crop HANPP formula to calculate the NPP harvested during timber harvest. Hardwood and softwood vary, so we adapted the formula to reflect their differences. This was done by using specific estimates from literature to estimate the different dry fraction, carbon content, and percent shoot of hardwood and softwood as for crops, as for crops. We derived timber HANPP measures for each county in each year: total metric kilotonnes of carbon and county-average density in gCm-2yr-1. We estimated grazing HANPP on public and private lands using data from USDA Forest Service, the Bureau of Land Management, and Landsat. For public lands, we started with raw data in the form of animal-unit months (AUMs) which are conceptually identical to HANPP. We converted AUM data to metric tonnes of carbon and on-site densities in gCm-2yr-1. For private lands we quantified the NPP contained in grassland and pasture in each county by matching 30m pixels classified as grassland/pasture with NPP data. We then categorized counties into USDA Land Resource Regions (LRR) that vary in the percentage of NPP utilized by livestock. We obtained data on the number of beef cattle per county. Total grazing for counties in each LRR was compared to NPP resources in those counties to derive a percentage of NPP grazed as the measure of HANPP. Grazing HANPP in each county was divided by the area of grassland/pasture to obtain onsite density and by the area of the county to obtain county-wide HANPP density in gCm-2yr-1. Crop, timber, and grazing HANPP were summed to get a measure of total HANPP for all counties. County HANPP was summed to get national HANPP. Calculations were done with Microsoft Excel and ArcGIS Pro software.
National Science Foundation