Recipient Organization
PENNSYLVANIA STATE UNIVERSITY
208 MUELLER LABORATORY
UNIVERSITY PARK,PA 16802
Performing Department
Ecosystem Science & Management
Non Technical Summary
Abrupt changes in a forest ecosystem, whether natural or anthropogenic, are changes that occur over short time periods; such disturbance has the potential to drive state changes and alter forest resilience. Understanding how present-day abrupt forest change may alter ecosystem services is becoming more important due to ever-growing anthropogenic stresses. Forest managers trying the adapt to anthropogenic stress can benefit from the study and quantification of past abrupt changes in forests, especially when the legacy of past disturbance is still evident. Across the United Kingdom, Europe, and recently the northeastern United States, the examination of historic forest change due to charcoal manufacturing for the firing of iron or lime furnaces is yielding new insights relative to landscape stability, anthropogenic vs natural soil genesis, and forest evolution. Our proposal strives to evaluate how historic land clearing for the charcoal industry (supporting iron furnaces) may have driven present day forest composition. We build upon three existing collaborations: Forest-deer interaction research with the Pennsylvania Bureau of Forestry; ecosystem evolution research with the US Department of Agriculture (USDA), National Resources Conservation Service (NRCS); and relict charcoal hearth (RCH) legacy effects research on forests with Brandenburg University of Technology Cottbus-Senftenberg, Germany. Our research focuses on two regions of historic charcoal production, the greater Greenwood State Forest area of Rothrock State Forest and the Michaux State Forest. We incorporate field sampling of forest vegetation (overstory and understory), hearth soils, and hydrologic parameters (in hearth and non-hearth areas) in order to quantify the uniqueness of relict charcoal hearth (RCH) systems. We will use the LANDIS-II forest landscape model to simulate forest composition with and without RCH production, and in relation to simulated timber harvests and deer browsing pressure. There is a tremendous gap in the knowledge regarding the general ecological significance of RCHs as we are just starting to realize the greater dimensions of large RCH landscapes that have recently been identified. Further research on RCHs is needed to enhance our understanding of the environmental consequences of historical charcoal production and examine the quantity and quality of this "legacy effect" on modern ecosystems. Deliverables from this proposal include: detailed site data on species composition and soils as affected by RCH development; support of continued regional, national, and international collaborations; training of a graduate student; numerous peer-reviewed publications; submission of proposals to external funding bodies; and the generation of one-of-a-kind historic landscape change knowledge benefiting state agencies, historians, and the landscape analysis community.
Animal Health Component
30%
Research Effort Categories
Basic
70%
Applied
30%
Developmental
(N/A)
Goals / Objectives
This proposal's objectives are: (1) to understand the site-specific effect that charcoal manufacturing had on the distribution, development and properties of forest vegetation and soils; and (2) to model the effect through time of charcoal manufacturing, and differing deer and forest management goals, on forests.
Project Methods
Task 1- Documenting the spatial extent of RCH production. (Led: Drohan, graduate student).(Deliverable 1) Maps of RCH and furnace occurrence. RCH mapping has been ongoing in study areas using statewide LiDAR data. RCH mapping and morphometry metric estimation is being conducted per Ludemann (2010), Raab et al. (2016), and Johnson and Ouimet (2018). Additional data on historic structures/roads is being provided by DCNR and PA State Museum staff. Soil auger surveys of hearths are being used to gather hearth metrics (Hirsch et al., 2017) for Task 2.(Deliverables 2 and 3) Statistics on RCH occurrence and morphometrics. Description of RCH size (mean, median, mode), distance to current and historic roads, distance to furnace, landscape position, elevation, and aspect will be recorded in a GIS (ESRI ArcMap 10.6). Per detailed inventory data from archeologists affiliated with each forest and our record searches, we will try to record hearth locations in time to ascertain periods of construction. PA DCNR stand mapping data is being used to ascertain the current stand type where RCHs occur.(Deliverable 4) Quantify the potential interaction between RCHs and landscape level hydrology. Per Shore et al. (2013), we will use the Network Index (NI) approach of Lane et al. (2004, 2009) to characterize landscape hydrologic connectivity in relation to RCHs.Statistical Approach: Analysis strategies will follow those of Hirsch et al. (2017), Johnson and Ouimet (2014), Raab et al. (2017), and Schmidt et al. (2016). Statistics will use a variety of univariate and multivariate methods based off R-packages (R-Core Team, 2019). Spatial analysis approaches will use derived RCH landscape patterns to train a Random Forest classification model in R, which will be parameterized with landscape predictors and associated RCH features (Hackkenberg et al., 2018; Liaw and Wiener, 2002).Task 2 aim - Collection and characterization of field soils and vegetation in RCH and control areas. Led: Drohan, graduate student, co-PI McDill.(Deliverable 1) Across each study location, 20 hearths and control areas without hearths will be chosen across a range of the timespan of hearth usage. Detailed soil profile construction and description will occur on 10 RCH and controls with carbon stock sampling on the remaining 10 sites. Control areas of similar native soil (but no closer than 100m) will be excavated to 150 cm or refusal. Hearths will be trenched across their dominant axis per Hirsch et al. (2017). Profiles will be described (Schoenberger et al., 2012) and sampled by horizon (Soil Science Division Staff, 2017) or geoarcheological layer (Raab et al., 2017). Hearth reconstruction surveying will follow field protocols of Raab et al. (2017) and use a transit, level and grade rod. All features, layers or horizons will be recorded with ~50 MP resolution photography. RCH ground penetrating radar (400 mHz antennae) surveys will be run per Filzwieser et al. (2018).(Deliverable 2) Quantify soil organic contests, types, stocks and quality. All methods per Soil Survey Staff (2014), and for each horizon or layer, include: (i) SOM content by loss-on-ignition (LOI) at 550°C; (ii) total carbon (C) and total nitrogen (N) by dry combustion; (iii) stocks (per Stolt et al., 2010); (iv) wet sieving for granulometry in combination with size-dependent characterization of SOM contents; and (v) assessing the amount of pyrophosphate extractable C by acid hydrolysis (Oades et al., 1984; Rovira and Vallejo, 2002), density fractionation (Crow et al., 2007; Cerli et al., 2012), and FTIR per Bonhage et al. (2018).(Deliverable 3) Quantify soil physics (all methods per Soil Survey Staff (2014)). For each horizon or layer, texture will be determined from size fractions using pipette analysis with pre-treatments, bulk density by volume (core volume) or nuclear density gauge, and unsaturated hydraulic conductivity by borehole permeameter (Aardvark brand). Thin section analysis will be used to quantify RCH micromorphology (XRF, SEM-EDX, XRF, Electron Probe).(Deliverable 4) Quantify soil chemistry and mineralogy (all methods per Soil Survey Staff (2014)). Soil pH will be determined in water and CaCl2; clay mineralogy and Fe and Mn oxide specific mineralogy will be determined via x-ray diffraction and operationally defined fractions (pyrophosphate and citrate dithionite bicarbonate); base cations, Al and Mn via an NH4Cl extraction (mechanical vacuum extractor); and total elemental analysis per aqua regia digestion.Statistical Approach: Analysis strategies will follow those of Hirsch et al. (2017) and Raab et al. (2017). Statistics comparing RCH and control soils will use a variety of univariate and multivariate methods based off R-packages (R-Core Team, 2019), including the Algorithms for Quantitative Pedology (Beaudette et al., 2013; GITHUB, 2019a).(Deliverable 5) Detailed species metrics. Per PA DCNR CFI mapping protocols (BOF-PA-DCNR, 2019), forest overstory and understory data will be collected on and off RCHs. In addition, understory plants on and around RCHs will be collected per protocols for the "Deer-Forest Study" following (Begley-Miller, 2018). Rare or threatened species will be noted in surveys per comparison via state and federal databases (BOF-PA-DCNR, 2019).Statistical Approach: Analysis strategies will follow those of Begley-Miller et al. (2019). Briefly, we will determine which covariates most explain occupancy across the study conditions for plant taxa using single-species occupancy models (MacKenzie et al. 2002) in the program MARK (White and Burnham 1999). For each taxon, we will use single-species occupancy models to estimate detection and occupancy probability, and odds ratios (Szumilas 2010).Task 3 aim -LANDIS-II models of RCH development, deer browse and forest management. Led: Drohan, co-PIs Diefenbach and McDill, graduate student.LANDIS-II initial communities mapping, management areas, and harvest prescriptions have already been developed by PI-Drohan for his statewide cumulative shale-gas impacts study. Browse extension metrics, relative to Pennsylvania forests, have also been developed via De Jager et al. (2017). (Deliverable 1) LANDIS-II modelling of RCH/no RCH development. LANDIS-II will run on model landscapes with and without RCH development. Clearing to build hearths will be modeled in LANDIS-II as a new forest prescription with recurrence cutting of 40-year rotations. Recurrence cutting has been determined from our work with Greenwood Furnace State Park education specialist Paul Fagley, Straka (2014), and Raab et al. (2017). (Deliverable 2) Quantification of RCH development versus no RCH development, with differing levels of deer browse (current versus lower Deer Management Assistance Program target levels), and current DCNR forest management goals as identified in the "Deer-Forest Study".Statistical Approach: Analysis strategies will follow those of De Jager et al. (2017), Pauli et al. (2015), Scheller et al. (2008). Briefly, all potential scenarios across the three primary variables will be compared [RCH (yes/no); Browse (status quo versus reduced deer population); and Harvest (status quo prescriptions versus 2019 DCNR forest plan goals)]. A mixed-effects analysis of variance will be conducted for each model using the lme4 package (Bates et al. 2013) in R (R-Core Team, 2019). The relative magnitude of effects (White et al. 2013) and comparison between simulated scenarios (Grimm and Railsback, 2005; Scheller and Mladenoff, 2008) will be applied rather than their statistical significance (White et al., 2013).