The original plantings of several Great Plains shelterbelts were part of the Prairie States Forestry Project (PSFP), one of the largest afforestation projects in US history. The federal project was part of the American New Deal, which began when Franklin D. Roosevelt (FDR) took office as president of the United States in 1933. As president, FDR established a series of experimental programs and projects to address unemployment issues faced by Americans during the Great Depression (1929–1939). During the Depression, Great Plains farmers were devastated by a combination of bank failure (resulting in loss of savings), extreme environmental drought, and suffocating dust storms, which rendered their land useless (McLeman et al. 2014, 418) (Fig. 1). This ‘Dust Bowl’ period impacted millions of acres of farmland, resulting in ‘distress migration on scales not previously seen’ (McLeman et al. 2014, 418).
In response to economic and environmental crises, FDR instituted a series of programs and projects aimed at addressing soil conservation and introducing new farming techniques across the region. During his first 100 days, FDR established the Civilian Conservation Corps (CCC) and the Soil Erosion Service, which later became the Soil Conservation Service, known today as the Natural Resource Conservation Service. Historian Woolner (2010) noted in his essay, ‘FDR and the New Deal Response to an Environmental Catastrophe’, that ‘the establishment of the Soil Erosion Service marks the first major federal commitment to the preservation of natural resources in private hands’. As part of federal soil conservation efforts, the experimental PSFP shelterbelt planting program was implemented in 1935. The size and scope were ambitious, and the federal project was achieved through cooperative planting agreements between the United States Forest Service and private landowners: from 1935 to 1942, the PSFP planted over 200 million trees in approximately 33,000 shelterbelts on private land across 18,599 miles (Perry 1942, 1) (Figs. 2 and 3).
The PSFP was managed originally by the US Forest Service, but by 1938 the Secretary of Agriculture, Henry A. Wallace (1933–1940), began to phase out the Forest Service’s involvement with the project and shifted responsibility for ‘farm forestry function’ into the conservation programs of the Soil Conservation Service (Droze 1977, 219). By 1942, one-fourth of the shelterbelt zone had been organised into soil conservation districts. As the United States entered World War II (1941–1945), a labour shortage and decreased financial support provided by the Soil Conservation Service resulted in a decline in shelterbelt planting as a conservation strategy across the Great Plains (Droze 1977, 227). A comprehensive and coordinated shelterbelt monitoring effort by the United States Soil Conservation Service did not occur after the project’s transfer from the United States Forest Service, making identifying the location of existing PSFP shelterbelts challenging for cultural and natural resource managers. Until recently, only anecdotal evidence existed indicating that many Great Plains shelterbelts, including those originally planted in the PSFP, have been removed from the landscape due to increased crop prices, new farming techniques, and age-related deterioration (Liknes et al. 2017, 167).
The PSFP headquarters were located in Lincoln, Nebraska, and original archival materials remain unpreserved at the USDA National Agroforestry Center (NAC) on the University of Nebraska-Lincoln’s (UNL) campus. In 1940, then project director Paul Roberts wrote a letter to all participating Forest Service state directors and division chiefs encouraging them to keep a permanent record of the project, saying, ‘You may not attach much importance to some of them, but send them in any way, for out of all of this someday someone will write a saga of the Shelterbelt. (Roberts, 1940)’ Robert’s efforts resulted in the NAC collection retaining documents for the three participating northern states, Nebraska, South Dakota, and North Dakota, with limited documentation for the southern project states of Texas, Kansas, and Oklahoma. The archival collection covers 1935–1942 and includes primary source textual material, including correspondence and cooperative planting agreements between the Forest Service project administers and private landowners referred to as cooperators. The US Forest Service cartographically recorded the PSFP using public land survey maps, placing them in the NAC archive. Additionally, each cooperator’s shelterbelt planting was supported by descriptive records, which included information on the quantity of species planted as well as the length and width of each shelterbelt. The documents are located in filing cabinets organised by participating states. Nebraska, for example, has 150 county maps, 1000 township maps, and over 6000 descriptive records.
Photography is also included in the collection, with over 6000 printed photographs taken by project administers and New Deal photographers hired to document the progress of the federally funded project (Fig. 4). Several duplicate photographs are stored in filing cabinets and photo albums. Photographs are labeled with an archival number that can be traced to negatives stored in the National Archives in Washington, DC, the date the photograph was taken, and the name of the photographer. The location is typically listed as a state or county with a supporting description of the planting site and the name of the cooperator if included in the picture. Additionally, photographs were not organised in the archive by location but by classification tags: shelterbelt, utilisation, damage, field windbreaks, crop protection, and human interest. The thematic archival photographic classification system makes identifying and spatially locating photographs challenging.
One important archival record that was identified for our cartographic purposes was the ‘Township Maps’ (Fig. 5). The archival township maps contained colour-coded plantings for all eight years of the project as well as the transcribed original landowner’s name to farm number. These maps became the basis for generating spatial data. The combination of location, attribute data, and planting dates in the archival sources provided the foundation for a data capture process in GIS.
Through the historical reconstruction of the PSFP, it is possible to identify the current condition of today’s remaining shelterbelts. For this research, historical sources (maps, records, and photographs) can reconstruct the history of this large-scale cultural landscape using GIS technology. The methodological goal is to apply what historian Anne Knowles calls a ‘historical GIS’ approach, combining archival research with modern GIS technology to document cultural landscape change (Knowles 2005, 7). Historical GIS combines methods from historical geography with spatial and digital history to address geographic questions associated with time and place (Knowles and Hillier 2008, 7). The majority of scholarship in this area presents historical evidence as a geographic argument in visual forms, such as thematic maps (Knowles and Hillier 2008, 7). In their book, Placing History: How Maps, Spatial Data, and GIS are Changing Historical Scholarship, Knowles and Hillier identify three areas of historical GIS scholarship:
One is empirical research on the history of land use and the development of spatial economies. The second is using GIS to visualise past landscapes and the changing morphology of the built environment over time. The third consists of infrastructure projects that aim to make historical source materials available to scholars and the general public for mapping and other kinds of analysis. (8)
This project contributes to Knowles and Hillier’s second and third areas of scholarship, ‘reconstructing past landscapes’ and ‘infrastructure projects’ by applying a historical GIS method referred to by geographer Gregory (2003, 103) as ‘data capture’ in which data from archival maps are translated into a publicly accessible GIS dataset. In addition to scholarly research, a novel approach for historical GIS methods can be applied by individuals or agencies working to preserve and provide stewardship for cultural landscapes (Donahue 2008, 174). In the context of the PSFP, the historical GIS method serves not only as a research tool but can also identify and evaluate opportunities for future preservation and stewardship of the PSFP shelterbelts.
For this investigation, Nebraska was selected as the study area, and 1000 archival township maps that include over 10,000 planted shelterbelts were manually digitised. Geo-referenced vector data was downloaded from the Natural Resource Conservation Service (NRCS) and formatted in ArcGIS as additional layers for the mapping process. Raster data was included from the National Agricultural Image Program (NAIP), which acquires and disseminates aerial images during the agricultural growing season. Aerial photography is taken by NAIP (2016) on a three- to five-year cycle, and aerial GeoTIFF from 2016 were downloaded as digital ortho quarter quad tiles (DOQO) for this study (Table 1).
Vector database with attribute data
A vector database was created from the historical township map and entered manually as a line in ArcGIS. Because the archival township maps were drafted to scale and descriptive data for length was available for each shelterbelt planted, manual entry was possible. Shelterbelts are typically quantified by miles planted (length) and provided the rationale for recording the belts as lines instead of polygons representing the area. Once each shelterbelt feature was digitised as a line, appropriate attribute data was added from the descriptions on the township maps: cooperator’s (landowner’s) name, year of planting, and legal description of the location, including farm identification number (Table 2).
An alternative process, described by Rumsey and Williams (2002, 5) as ‘rubber sheeting’, allows one to overlay maps from different eras and adjust the transparency of the layers until the evidence of change in the maps is visible in GIS. The rubber sheeting process, however, would have required additional steps in the process, including the scanning of thousands of archival documents. Therefore, we decided to enter the length of each shelterbelt planted manually.
Overlay analysis using raster data
The historical shelterbelt vector layers were overlaid on current aerial geo-data, making it possible to analyse and visualise remaining shelterbelts in GIS (Fig. 6). Three categories were established to classify each shelterbelt’s existing condition in 2016 (removed shelterbelts, partially intact shelterbelts, and fully intact shelterbelts). The condition of the shelterbelt was analysed based on visual evidence matching the archival documents of the remaining historical planting patterns (length and continuity of planting) resulting in continual land use (performance) as an agricultural conservation measure.
A fully intact shelterbelt retained over 75% of its original planting structure, as evident in the aerial photography, maintained original length and appropriate continuity, and continues to function as a wind buffer. Partially intact shelterbelts were belts that retained 6% to 74% of their original planting structure, with the goal of inventorying as many partially intact plantings as possible. A partially intact shelterbelt, for example, could have maintained several rows in width but had significant gaps in length that could compromise the original designed land use and performance. Additional research would be necessary to determine which partially intact shelterbelts continue to function as a conservation measure. If less than 5% of a shelterbelt’s planting structure remained, the planting was classified as removed. Additional fields, including a GIS-designated reference number, were added to the attribute data, classifying each planting as fully intact, partially intact, or removed (Table 2).