The Road to Pierce Farm Connecting Land-Use History and Soil Phosphate Concentration David Diaz Harvard University Pierce Farm: Mapping Topography: Mapping Soil Phosphate: Grid Phosphate Mapping Soil Phosphate: (0-10cm) Interpolated Phosphate Introducing Geography Digital Elevation and Coring Grids 0 25 50 100 150 200 Meters Interpolating Cores to Historical (0-10cm) Pierce Farm is situated on the Prospect Hill Tract of Harvard Forest, an area which Contour Lines Examination of soil phosphate distribution by grid reveals increasing levels of soil ( ( ( ( ¯ Legend Land-Use 0 25 50 100 150 200 Meters ¯ includes significant topographic variation as phosphate with increasing intensity of Soil Cores (0-10cm Data) When the soil cores were mapped to their respective land-use ¯ mg P/kg soil well as distinct land-use histories. The two of historical land-use which the grids were ( 1 polygons instead of intended land-use type grids, a different ( 10 these factors cannot easily be separated in the intended to represent. Within the grid ( ( ( 100 picture appears. It should first be noted that the potential error (( (( Historical Land-Use present or the past because historical land- framework, both improved pasture and Occupied of GPS data points may confound the joining of these two Legend use was contingent upon geographic qualities cultivated field have comparable phosphate Cultivated Mowing/Improved features near boundaries of land-use polygons, explicitly at the Soil Cores (0-10cm Data) Phosphate (0-10cm) such as drainage and accessibility. levels. While phosphate levels appear to be Unimproved Pasture southern boundary of the Improved Pasture area. This map mg P/kg soil Woodlot approximately the same between 0-10cm and ( ( Phosphate (0-10cm) may be contrasted with that above with the critical point that 33 There is also a large historical influence in mg P/kg soil some of the cores within a particular land-use coring grid are 53 15-25cm for both Woodlot and Unimproved ( ( 15 107 understanding land-use from generation to ((( 33 not always within the same land-use polygon. 146 Pasture, phosphate concentration increases (( 242 generation as conversion to more intense land ( 57 greatly with depth in Improved Pasture, ((( 102 use requires progressively more work on the In contrast to the grid phosphate data, at 0-10cm, Improved (( ( 107 Cultivated, and Occupied Grids, nearly ( 242 land, from clear-cutting woodlot for pasture to (( Pasture surpasses Cultivated in soil phosphate concentration doubling for Cultivated and Improved Pasture ( ( mowing unimproved pastures to increase the from 0-10cm to 15-25cm. These changes by (( (((( ( ( and the two are easily distinguished. At 15-25cm however, productivity of grasses. In the opposite 0 250 500 1,000 Meters this relationship is reversed and the two are still easily depth seem to indicate that historical soil distinguished. At a depth of 0-10cm, the phosphate levels of direction, when plots begin to be abandoned in the late nineteenth century, land-use recedes The topography of Prospect Hill has been disturbance may initiate some type of Grid Phosphate Unimproved Pasture are relatively close to that of Woodlot. feedback cycle which acts upon these two soil (15-25cm) This relative similarity disappears at a depth of 15-25cm as to the most desirable and productive areas, shown in this map by layering a historic depths at different rates. 0 25 50 100 150 200 soil phosphate in Unimproved Pasture increases substantially predominantly centering on cultivated areas USGS Contour Map over a modern digital Meters and homesteads. An important shortcoming in using soil disturbance profiles to reconstruct elevation model which has been hillshaded to create the optical illusion of three- Within each grid the phosphate levels of each core can be seen to vary widely. The ( ( ( ( ¯ Legend Soil Cores (15-25cm Data) with depth. These maps were created with a spatial join between soil Interpolated Phosphate (15-25cm) 0 25 50 100 150 200 ¯ mg P/kg soil Meters historical land-use is that it can only detect the dimensional space. The Pierce Farm westernmost grid has suspiciously low ( 1 cores and historical land-use polygons. The hue of the most intense land-use, preventing us from (highlighted with the red outline above) is phosphate levels considering it is assumed to ( 10 polygons indicating phosphorus levels is representative, witnessing this retreat. situated on a plot incorporating the lowland swamp area at the base of Prospect Hill be the same land-use as the grid 150m to the northeast. This may be due to the low ( (( ( ( ( ( 1,000 meaning it closely portrays the contrast of phosphate levels Reconstruction of this retreat and Historical Land-Use between polygons with a synchronous contrast in hue. moving southwest away from the summit sample size of the westernmost grid, but Occupied abandonment may be accomplished through Cultivated The extraction of more soil cores would facilitate a more upwards in elevation back towards the merits further investigation. Mowing/Improved Legend the historical record and be explained in Unimproved Pasture accurate interpolation of soil phosphate, particularly along Soil Cores (15-25cm Data) town road and later administration ( ( Woodlot Phosphate (15-25cm) relation to present-day forest cover and These maps were created with a spatial join Phosphate (15-25cm) boundary zones, or with the intent of discovering boundary buildings of the Harvard Forest. ( ( mg P/kg soil undergrowth (Foster 1992, Motzkin et al. between soil cores and the soil coring grid ( mg P/kg soil zones through soil phosphate alone. 23 (( ((( 33 108 1999). Swamp areas can be seen on from the polygons. The hue of the grid indicating 38 163 (( ( 89 195 In the present, the wayward visitor to the historic USGS map to occupy most of the phosphorus levels is representative, meaning ( (( (( 175 196 303 low-lying elevations on the Prospect Hill it closely portrays of the contrast of phosphate ( woods may find stone walls and a cellar hole 303 of the Pierce Farm, but would likely miss a Tract. In contrast to neighboring Prospect levels between grids with a synchronous (( ( Hill, Pierce Farm encloses relatively contrast in hue. Coring grids were established large part of the history of the region without smaller slopes, likely providing for different in the field in 80m x 20m rectangles, intended knowing what has happened beneath the drainage patterns and perhaps more to cover local variations in slope and surface of the land for the past two and a half productive soil than the land parcels to the vegetation. They have been reconstructed centuries. north across the town border with here as 90m x 30m rectangles to incorporate The objective of this project is to introduce Phillipston. Despite the smoother slopes the potential error within GPS coordinates. what happened historically above and below of Pierce Farm in the higher end of the Note: the Occupied Land-Use polygon is not the surface of Pierce Farm and relate that plot, the presence of the swamp at the visible because it is the exact same shape as history to present soil conditions. base of Prospect Hill in the northeastern the Occupied Coring Grid. region of the Farm surely had an effect on historical land-use. This swamp region is documented by the Harvard Forest not as Methods: Woodlot, but instead as Unimproved References: Pasture, suggesting that despite its soggy Displaying Geography ground and seemingly undesirable qualities, it may in fact have been utilized Foster, David R. “Land-use history (1730-1990) and vegetation dynamics in central New Through the utilization of Geographic by the occupants of Pierce Farm. England, USA.” Journal of Ecology 80 (1992): 753-772. Information Systems, the geography of the land may be displayed and analyzed in Motzkin, Glenn et al. “Vegetation patterns in heterogeneous landscapes: The relation to several different elements. The importance of history and environment.” Journal of Vegetation Science 10 (1999): spatial framework of GIS facilitates visual 903-920. comprehension of multiple variables and relationships and allows for a more holistic appreciation of spatial relationships. Conclusion: GIS Data: While GIS analysis may provide incredible potential for manipulating and displaying Historical Land-Use polygons modified from existing Harvard Forest data. The materials used to construct these maps vary in source, context and date of initial spatial information, it is contingent upon the accuracy and precision of the underlying spatial data. The source of the most confounding errors is the GPS unit used to The 1946 USGS Map was taken from University of New Hampshire Government construction, but may be fused together to Documents Department Historic USGS Maps of New England and New York website. offer a specifically tailored spatial portrayal of determine the UTM coordinates of each soil core. At such a small scale, errors of up to 10m can become increasingly difficult to reconcile with the grid on the ground as well as (Available at http://docs.unh.edu/nhtopos/nhtopos.htm) the Pierce Farm area and provide fertile ground for interpreting changes in the land. complicate any spatial joins and interpolation between soil cores and land-use layers. USGS aerial photos of the Prospect Hill tract in Harvard Forest collected from The data for this mapping project are The use of GIS technology has enabled the interpolation of soil phosphate values and TerraServer. (Available at http://terraserver.microsoft.com) displayed primarily through the use of Vector- allowed the visualization of contrast between historical land-use regions based on soil Relational methods as opposed to Raster phosphate concentrations. This spatial analysis highlights a pattern of increasing soil based GIS applications. All of these maps phosphate with corresponding increases in historical land-use intensity. displayed in UTM 18N projection The most valuable aspect for GIS in interpretation these experimental data is in the recognition of spatial error and shortcomings of both GIS datalayers as well as on-the- Acknowledgements: ground experimental design. The boundaries created in GIS layers are not necessarily the same boundaries that can be seen in the forest composition and soils, nor should Special thanks to Noreen Tuross, Joshua Wright and Cheryl Makarewicz for leading the they be misconstrued to be error-proof. In this circumstance, GIS analysis has research project, creating positive and encouraging feedback, and providing highlighted the problem with data collection at boundaries where precise measurements transportation to the site and to the apple orchard, among other things. of geographic location were not available. The integration of GIS into experimental design may facilitate more productive Thanks to Dave Orwig and Paul Cote for their help in creating and “wrangling” this data distinctions between boundary zones and enable more definitive results. and to Bronwen Everill for her help conducting research.
Pages to are hidden for
"Historical Land Research"Please download to view full document