Progress 07/01/15 to 06/30/17
Outputs
Target Audience:The target audience for this research is groups involved in land conservation from both the public and private sectors. We dissemminated our results to a diverse and highly interdisciplinary audience during this reporting period. Presentation of results from this study ocurred at a national conference of Geographers within a session focused on land change analysis. This audience included both physical and social scientists engaged in studying impacts of land change. Since results from the Year 1 analysis also formed the basis for some collaborative work within students and faculty from the Rutgers Graduate Program in Ecology, results were also presented as part of two other professional conferences by a graduate student from that program. Audience members included ecologists, conservation biologists, professional planners, and land management professionals. All findings are currently being incorporated into draft manuscripts that will be sent out to peer-reviewed journals in 2018. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?The project has provided support for two years in both the training ad professional development of a PhD graduate student. This included training in both the theory of statistics and its practical application via programming software packages like Stata and R. The graduate student also received training in the theory and use of spatial analytical software and relational database management. Professional development opportunities included taking the lead in both grant and report writing for the project, preparation of manuscripts for publication, and preparing and presenting both posters and oral presentations for regional and national conferences. How have the results been disseminated to communities of interest?During the second reporting period (October 2016 - October 2017), we shared Year 1 findings through a major national conference, the American Association of Geographers (AAG). This is a conference of academic scholars and GIS professional with an interdisciplinary approach to problem solving. Attendees at the conference include people working in both the social sciences and physical sciences. As such, it targets a different and much wider audience than the conferences attended in Fall of 2016. Dissemination of findings was also supported through preparation of a PhD dissertation. The dissertation is composed of three draft manuscripts that will be sent out for review in 2018 to journals with a focus in ecology, geography, and land management. What do you plan to do during the next reporting period to accomplish the goals?
Nothing Reported
Impacts
What was accomplished under these goals?
Sandwiched between several major metropolitan regions, the forests of New Jersey's Highlands region have experienced a disproportionate amount of urban expansion, but also an impressive amount of land conservation. Selection of conservation land there, as throughout the U.S., is highly opportunistic and often relies upon a landowner's choice to make that land 'available' for conservation. Consequently, it is important to understand whether conservation networks assembled within these constraints have spatial characteristics desirable for sustaining existing species and processes. Three measurements of interest in answering this question are 1) the relative size of habitat fragments, 2) the relative land change around those fragments, and 3) the relative conservation value of recently protected lands. We found that 1) roughly a third of forest land (both protected and unprotected) occurs in fragments 100-500 hectares in size, 2) urban land has increased roughly 10% on the borders of Protected Forests in the last 26 years, 3) from 2000-~2011 high conservation lands had a lower rate of urban conversion than low conservation lands, and conservation efforts were proactive rather than reactive. Taken together, these results suggest that opportunistic acquisition practices have enabled protection of some large and high quality tracts of forestland. However, because the largest acquisitions have occurred in regions less desirable for development, there is a spatial bias in the lands which are protected. Regions in demand for development tend to have smaller conservation tracts and these will potentially be more compromised by land change on their borders. Goals 1 & 2 The analysis of land change surrounding PFs was completed in Year 1. Key findings include documenting that forest habitat fragments 100-500 hectares (ha) were the most common (by count) and most abundant (by total area) of all forest fragment size classes across the NJ Highlands. This is also the most common and most abundant size class of forest patches formally designated as Protected Forest (PFs). Because of this, we recommend that both academics and practitioners prioritize the development of management and restoration practices for fragments of this size. An understanding of forest community dynamics at this scale will be applicable to managing about 32% of the existing forest cover. Our analysis of land change in the 250 meter buffer surrounding each PF fragment showed loss of agricultural land and increase in urban/developed land for all size classes. Although these aggregate patterns of land change have been documented across the Highlands, it was previously unclear the nature and rate of dynamic land change in the buffers around Protected Forests. Our analysis documents that land change is occurring in close proximity to PFs, and that the rate of urban increase is ~10% per year on average for forest fragments less than 500 hectares. Goals 3. The analysis for this goal has been the focus of Year 2 activities and builds upon work from a prior analysis conducted within the Highlands. As part of collaborative report by the US Forest Service, GIS layers were developed to assess and map the relative conservation value of land which were uncommitted to either protection or urban status. Conservation value was assessed with five criteria: forest quality, agricultural quality, recreational value, biodiversity value, and water protection. The result was a composite GIS layer mapping the remaining high value and low value lands in 1999. Combining this layer and the Protected Area database (developed as part of our Year 1 analysis), we calculated the total amount of low and high quality land which transitioned into protection from 2000-~2012. There were an estimated 50,027 hectares of low value lands across the Highlands in 1999. Our analysis showed that these lands had probabilities of 0.09 for transitioning into protection, 0.12 for transitioning into urban, and 0.79 for not transitioning (i.e. being "uncommitted"). There were an estimated 103,560 hectares of high value lands across the Highlands in 1999. These lands had a probability of 0.19 for transitioning into protection, 0.06 for transitioning into urban, and 0.75 for not transitioning. Thus, although the total amount of land converted to urban from both the low and high value land pool was similar (5766 ha, vs 5956 ha, respectively), the probability of urban conversion was lower for high value lands. This is an important finding as the lower conversion rates suggests that development patterns are "sparing" some high value conservation lands. This study also offers an important baseline for any future policy and advocacy efforts aimed at further decreasing the conversion rate of high value lands. Goal 4 We focused on modeling the influence of one specific type of land change (urban development) relative to conservation acquisitions within the Highlands. We used regression to model the correlation between development pressure in each municipality prior to 1999 and the amount of high value land which a municipality protected from 2000-2012. Specifically, we asked whether the most land conservation occurred in municipalities with more development pressure prior to 1999, or whether the most land conservation occurred in municipalities where there was an abundance of natural lands prior to 1999. The former would characterize a "reactive" approach to land conservation. (That is, protection occurred as a reaction to development). The latter would characterize a "proactive" approach to land conservation. (That is, land was protected before it became highly valuable for development.) We calculated six variables for each of the 83 municipalities. Two of the variables quantified the amount of natural land. The other four variables were metrics of development pressure. Specifically, the six variables were: 1. hectares of high value uncommitted land, 2. hectares of protected land, 3. change in population density, 4. increase in impervious surface, 5. increase in mean housing price, and 6. increase in average per capita income. The best model (adjusted R2 of 0.84) showed that high amounts of uncommitted and protected land by 1999 were the strongest predictors for greater amounts of land protection in a municipality from 2000-2012. However, two variables for urban development were also significant predictors. The best model showed that as impervious surface and income rise, there is a decrease in the amount of high value land protected. These results suggest that land conservation across the region is primarily characterized by a proactive approach. That is, the most land is conserved in regions with an abundance of natural lands and less development pressure. This approach has obvious benefits for biodiversity protection because large areas are protected before becoming threatened by development. However, it also creates biases in the selection process. Biodiversity on lands desirable for development will not get protection if they are too economically or politically costly to conserve. Goal 5. The data are currently stored as personal geodatabases in ArcGIS 10.2 and managed as a relational database in MSAccess software. Last year, we reported that the relational database management (rdbm) activities would be transitioned to PostgreSQL. The planned upgrade to ArcGIS 10.3 (which would facilitate such a transition) did not occur due to logistical conflicts. The database management transition was not a part of the original project proposal, so all original project objectives have been met with the current database. We still hope to transition rdbm activities into PostgreSQL software, but the timing and version of an ArcGIS upgrade is currently unclear.
Publications
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2017
Citation:
"Land Change around protected forests of the New Jersey Highlands: 1986-2012". Katherine Fenn. The American Association of Geographers Annual Meeting. 2017, April, 2017. Boston, MA.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2017
Citation:
How Small is Too Small? The Role of Size and Surrounding Landscape on Bird Communities in an Urban Landscape . Jeffrey Brown, Julie. L Lockwood, Katherine Fenn. The Ecological Society of America Annual Meeting. August, 2017. Portland, WA.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2017
Citation:
How Small is Too Small? Why the Size and Surrounding Landscape Matters when Planning Protected Spaces . Jeffrey Brown, Julie. L Lockwood, Katherine Fenn. International Urban Wildlife Conference. June, 2017. San Diego State University, CA.
|
Progress 10/01/15 to 09/30/16
Outputs
Target Audience: The target audience for our research includes groups involved directly and peripherally in land conservation both in the private and public sector. During our first year, we were able to reach conservation professions in both these sectors through presentations made at two different conferences. One conference targeted an international audience of conservation academics and practitioners. The second targeted New Jersey professionals engaged in planning and managing the built environment of the state. Our database of Protected Areas was also integrated with data from another McIntire-Stennis recipient for a collaborative publication on the economic aspects of Protected Area selection in urbanizing environments. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided? The funding has primarily supported the academic training of a PhD student through the Rutgers Geography Department. The project is interdisciplinary in nature, requiring analysis into policy history, ecological systems, population demographics, and technical mastery of GIS software. This type of multi-disciplinary training for graduate students also prepares them to work with a variety of conservation professional and to effectively communicate across disciplines in order to address challenges in natural resource management. How have the results been disseminated to communities of interest? The graduate student working under this grant attended two professional conferences this fall to present the Year 1 research results. The conferences were deliberately selected to target a diverse audience, the first being a group of New Jersey professionals engaged in planning and research within both the built environment and undisturbed landscapes (i.e. Mid-Atlantic Chapter of the Urban and Regional Information Systems Association). Members of the group work on a diverse array of projects, but use Geographic Information Systems as a common tool in their work. The second conference was an international student conservation conference which provided the opportunity to present results to peers, mentors, and professionals all engaged in the common practice of conservation science. What do you plan to do during the next reporting period to accomplish the goals?Goal 3 & 4. Funding for Year 2 is earmarked to address these goals. The data for these two analyses have been prepared. Although we had hoped to begin analysis in Fall 2016, preparation and dissemination of the results from Year 1 took priority during this period. We still expect to complete goals 3 & 4 in Spring 2017 and have results ready to report by June 2016, as this portion of the research is all computer-driven; there is no additional data collection required for these analyses. Goal 5. The data are currently stored as personal geodatabases in ArcGIS 10.2 and managed as a relational database through the Microsoft Access software. However over the course of the next year, pending institutional upgrades to ArcGIS 10.3, we intend to transition the relational database management activities into PostgreSQL software. (ArcGIS 10.3 and 10.4 offer easier communication between the Arc software and the Postgres software.) We will also identify GIS repositories appropriate for storing and making our data, processing methods, and metadata publically accessible. We will continue to disseminate Year 1 results in Spring of 2017. Presentations will target both academics and practitioners engaged in land and forest conservation. Results will be presented via a poster session at the New Jersey Land Rally in Spring 2017. This is an annual meeting of NGOs, government, and private individuals engaged in land protection within New Jersey. Results will also be presented at the American Association of Geographers (AAG), a national meeting of geographers.
Impacts
What was accomplished under these goals?
Land protection in New Jersey has been a highly success endeavor in the last two decades even as the state has experienced a simultaneous increase in the amount of urban land development. The consequence of these two competing land uses is a complex spatial interweaving of protected and urbanized lands across the landscape, the feedbacks between which are not well understood. These feedbacks are nonetheless pivotal to the long-term value of Protected Areas as refuges for biodiversity and sources of ecosystem services such as clean water. This research identifies the landscape patterns and spatial relationships between the built and protected environment. It is a first step toward helping land management agencies synthesize data about the human and ecological relationships, both of which must be considered as part of any long term management plans for Protected Areas. Our first year of research used spatial data analysis within a GIS to determine that fragments 100-500 ha were the most common and most abundant size class of forest (by total area) across the NJ Highlands. This is also the most common and most abundant size class of forest patches formally designated as Protected Forest (PFs). (About 32% of the total area of forest under protection currently occurs in patches of this size.) Because of this, we recommend that both academics and practitioners prioritize the development of management practices for fragments of this size. Our first year of research also analyzed land cover within a 250m buffer around PFs of all sizes. We found that the relative amount of urban land increased from 18% in 1986 to 23% in 2012. This increase in urban land proximate to PFs, and the attendant increase in the number and diversity of landowners associated with that urban land, may be detrimental or beneficial for PFs. The outcome will depend upon the way the nearby human population interacts with PFs. With an increase in landowners proximate to a PF may come an increase in the stakeholders with direct and indirect influence upon the management of that PF. Nearby landowners practices may introduce new ecological and human conflicts for PFs (e.g. through exotic species introductions and/or demands landowners may make on PF managers.) However, this urban increase also signals an opportunity to engage more citizens in the stewardship of PFs at deeper levels, potentially easing the burden upon land managers and improving the quality of management. Goal 1: The analysis of land change surrounding PFs is complete. Results are drawn from a Protected Forests database we assembled using four other Protected Areas databases, and the New Jersey Land Use Land Cover datasets for 1986 and 2012. After identifying the location of all Protected Forests (PFs) greater than 5 hectares (ha) (a total of 1287 PFs), we used GIS software to analyze them according to size class (5-10 ha, 10-25 ha, 25-50 ha, 50-100 ha, 100-500 ha, 500-1000 ha, 1000-5000 ha). We also analyzed the land cover composition within a 250 meter buffer of each PF. This buffer distance was selected because it covers the region in which land cover and landowner change may have a direct impact upon the course of management action. Buffers were then aggregated and analyzed according to PF size class. For each size class we calculated the relative percent of three land cover types (Agriculture, Urban, and Natural) in each buffer using land cover data from 1986 and 2012. Our results show that the land within 250 meters (m) of the borders of Highlands Protected forests has shifted in overall composition from 1986-2012 through loss of natural and agricultural land cover, and gains in urban land cover. The direction of these loss/gain patterns was consistent for all size classes of PFs, although the magnitude varied (as indicated by the standard deviations reported in the following). Relative percent urban land increase from a mean of 17.8% (± 3.2%) to 22.6% (± 4.8%). The relative percent of agricultural land decreased from a mean of 11.4% (± 7.8%) to 8.8% (± 5.9%). The relative percent of natural (i.e. forest and wetlands) decreased slightly from a mean of 70.8% (± 10.7%) to 68.6% (± 10.6%). Goal 2: To represent the rural-to-urban gradient of the buffers surrounding PFs, we created a thematic map of the PFs in ArcGIS 10.2. Themes or categories representing the type of land use along a rural-to-urban gradient were derived from a tripolar graph adapted from Riitters et al. (2009). The tripolar graph allows us to represent the relative abundance of the three dominant land covers (Agriculture, Urban, and Natural), the totals of which sum to 100% in every 250m buffer. Using a modified version of Riitters' graph, we determined the land cover category for the buffer of each PF. This categorical data was then passed back to the PF polygon in ArcGIS, and all PFs were colored according to the dominant type of land cover occurring within that buffer. This map provides a quick and easy to use visualization of land change occurring around the NJ Highlands Protected Forest network. From it, land managers can easily identify whether their forests have experienced recent land cover change on the boundaries, and what kind of change has occurred. Understanding land change dynamics will help in the interpretation of monitoring data collected within Protected Forests. Goal 3 and 4. Analysis for these goals is currently in progress and has been the focus of the Year 2 activities. Work to date on these goals have involved revisiting, supplementing, and cleaning up data on the year of establishment for individual the tax parcels which comprise the Protected Areas. We have also been reviewing the available data on spatial locations of biodiversity. With the data now mostly assembled, analysis on for this project will begin before the end of 2016. Goal 5. The first year of our project produced a database of Protected Areas covering the New Jersey Highlands ecoregion. We standardized Protected Area boundaries to the boundaries of the MOD IV tax parcel data layer produced by the State of New Jersey. Our data are therefore is linked to, and queryable by, that tax dataset. This increases the power of the Protected Areas database by integrating information on deeds, title, dates of purchase, and owner.
Publications
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2016
Citation:
Fenn, Katherine (2016, October). Connected, but not protected: Using GIS to map different amounts of land protection within forest fragments of the NJ Highlands. Poster session presented at the Mid-Atlantic Chapter of the Urban and Regional Information Systems Association, Atlantic City, NJ.
- Type:
Conference Papers and Presentations
Status:
Other
Year Published:
2016
Citation:
Fenn, Katherine (2016, October). Land change around protected forests of the New Jersey Highlands: 19862012. Presentation at the Student Conference on Conservation Science, New York, NY.
- Type:
Journal Articles
Status:
Published
Year Published:
2016
Citation:
Burkhalter, J. C., J. L. Lockwood, B. Maslo, K. H. Fenn, and K. Leu. 2016. Effects of cost metric on cost-effectiveness of protected-area network design in urban landscapes. Conservation Biology 30:403412.
|
Progress 07/01/15 to 09/30/15
Outputs
Target Audience:
Nothing Reported
Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest?
Nothing Reported
What do you plan to do during the next reporting period to accomplish the goals? The next reporting period will be devoted to completing Goals 1 and 2. To do this, we will be contacting tax agencies in select townships to try and identify missing acquisition years. It is not clear how much information return will be realized per unit effort. Since the location of the missing data impacts the scope of the analysis, our expectations for the analysis may be adjusted in future reports. (For example, parcels from Somerset County with missing acquisition years may be dropped from the analysis if it is too time-consuming to recover these data.) The data preparation and acquisition phase will end by spring. Analysis to address Goals 1 & 2 have already begun and will continue through the spring. During the next reporting period, we will also invest time in locating the GIS layer of priority acquisition sites from a 1992-1993 USDA-FS technical report. If found, we will review this and the other priority acquisition data layers (from the Landscape Project and the NJ Natural Heritage Program) to determine if all or a subset of these are appropriate for the Year 2 analysis. We expect to have the data layers needed for Goals 3-5 located and reviewed by the end of the next reporting period. This will keep us on target with our proposed Year 2 work.
Impacts
What was accomplished under these goals?
Impact Statement This study will inform the activities of land acquisition/conservation groups, and government agencies that provide tools and funding to support conservation efforts. New Jersey is reaching a pivotal point in terms of Protected Forests. The amount of forested land available for preservation is decreasing, causing land conservation groups to slowly shift their focus from active acquisition to active management of existing protected lands. Effective management will require an understanding of how high and low quality forests are assembled within the broader landscape of human land use. Summary data from this assessment on the land use dynamics surrounding PFs and their performance as PAs should help agencies prioritize management goals for individual properties. The second part of this study will benefit the state and federal agencies supporting land acquisition by describing trends in the quality and quantity of PFs selected for protection. Analyzing these trends will provide insight into how effectively open space planning tools have been used in the past, and suggest ways in which to support the future wave of management needs. New Jersey's relative success in preserving large, intact, forest parcels important to regional biodiversity can serve as a guide for other mid-Atlantic states trying to protect forests experiencing similar urban/suburban growth pressure. Specifically, it provides understanding of how effectively the current diffuse, multiagency approach to land preservation (one which is often constrained to select from a limited pool of 'available' land), addresses regional forest conservation and biodiversity goals. Project Activities All five objectives of this project rest upon having a spatial database of the protected areas located within the NJ Highlands. These spatial data are overlaid with land use/land cover data to identify the forested regions which are the focal "Protected Forests" in this study. Consequently, the first few months of this project have been devoted to integrating four spatial datasets which identify the spatial boundaries of Protected Area (PAs) in the NJ Highlands. During this reporting period, we acquired these data and standardized them to the spatial tax parcel dataset for each county. The output of this is a database of all tax parcels with protected status falling within the NJ Highlands portion of Bergen, Hunterdon, Morris, Passiac, Somerset, Sussex, and Warren County. These parcels will be matched to specific acquisition year contained from other data sources. After combining the datasets, we a binned tax parcels according to size ((0-0.1, 0.1-0.5, 0.5-1, 1-5, 5-10, or 10-50 kilometers square (km2)) and summed the area within each bin. Each area reported is rounded to the nearest km2. These data were of interest because PA acquisitions follow tax parcel boundaries, not ecological boundary (such as watersheds). Tax parcels offer an assessment of the degree to which land has been subdivided for human use before passing into protected status.Spatially standardizing PAs to tax parcel data also links them to land values and ownership records. According to our database, there are a total of 3525 individual tax parcels with protected status within the NJ Highlands (2508, 615, 152, 207, 20, and 23 parcels, respectively, in each size class listed above). Together these cover 1273 km2, roughly 33%, of the total area within the NJ Highland ecoregion. This 1273 km2 is distributed unevenly across the size classes (44 km2, 146 km2, 108km2, 413 km2, 138 km2, and 424km2 respectively, in each size class listed above). Our results show that the smallest size class has the largest number of individual parcels (2508), but represents only about 4% of the total area under protection. The largest size class has the second smallest count of parcels (23), but represents 33% of the total area under protection. Progress toward Objectives 1 & 2: We integrated our Protected Areas dataset with statewide land use/land cover data. We identified contiguous forested regions within the Protected Area using the 2012 land use/land cover data. These 'patches' of Protected Forests are the units of interest in our study. As above, we a binned forest patches according to size (0-0.1, 0.1-0.5, 0.5-1, 1-5, 5-10, or 10-50 kilometers square (km2)) and summed the area within each bin. Note that the boundaries of each forest patch in our analysis are defined by either non-forested land, or the presence of unprotected forested land. Each area reported is rounded to the nearest km2. Our analysis showed a total of 9191 individual forest patches within the Protected Areas of the NJ Highlands (88, 123, 85, 249, 129, and 116 patches, respectively, in each size class listed above). Together these cover 790 km2 (62%) of the total 1273 km2 of the NJ Highlands Protected Areas. Eleven percent of the total forest area exists in patches 0-0.1 km2, 15% exists in 0.1-0.5 km2 patches, 11% exists in 0.5-1 km2 patches, 32% exists in patches 1-5 km2, 16% exists in patches 5-10 km2, and 15% exists in patches 10-50 km2. Our preliminary results show the level of forest fragmentation within existing PAs. Although these patches may be connected to forest beyond the PA boundaries, roughly a third of protected forests currently exist in patches less than 1 km2, a third exists in patches 1-5 km2, and a third exist in patches 5-50 km2. Although the NJ Highlands Protected Forest network is still a work in progress, this analysis suggests the shape of the emerging network. Management of biodiversity and ecosystem services within these protected areas will be constrained by forest patch size. Progress toward Objective 3: During this reporting period we confirmed the presence of an early (1995) version of the Landscape Project available through the Rutgers Center for Remote Sensing and Spatial Analysis. Discussion with CRSSA personnel also revealed a 1992-1993 USDA-FS technical report which identifies priority sites for land acquisition in the NJ Highlands. We believe will we be able to access the GIS data layers used in the report, so we will be trying to find them in the upcoming months. Accessing this layer would allow the analysis to reach back a few more years. It also provides an alternative land priority ranking different from the Landscape Project, a spatial dataset which focuses most on habitat for endangered wildlife. Progress toward Objective 4: We have been exploring some analytical approaches to this question, including the use of a Bayesian framework for statistical analysis of transition probabilities. Progress toward Objective 5: We have been compiling the parcel-level acquisition dates which should be of interest to include in a state-level dataset.
Publications
|
|