1. Background and Research Question:

The goal of my work in this course is to assess the influence of forest governance on spatial patterns of forest disturbance. Forest governance can be understood as forest-related decisions, their implementation and resulting effects within a given institutional setting, whereas forest disturbances are events that cause change in the structure and composition of a forest ecosystem. For this course, I’ll be focusing my analysis on disturbances associated with timber production, e.g., clear-cutting and partial harvests. My study area is Willamette National Forest, which encompasses roughly 6,800 square km in the central portion of Oregon’s West Cascades.

Governance of the Willamette and other federally managed forests of the Pacific Northwest is shaped largely by the Northwest Forest Plan (NWFP). A key aspect of the NWFP is a system of land use designation (LUD) in which spatially explicit zones are managed according to a single or dominant management priority (Charnley, 2006). For example, wilderness areas are designated as “Congressionally Withdrawn” and are thus protected from timber harvest, while “Matrix Lands” are those on which timber harvest is concentrated. The various LUDs, along with interspersed state, private and other lands, form a mosaic of ownership and management priorities that are manifested as disturbance patterns in the forest landscape. And so, the research question I’ll be addressing is: How do land use designations and ownership influence patterns of disturbance in Willamette National Forest?

2. Datasets:

I’ll be relying primarily on a Landsat imagery time-series (30 meter spatial resolution, 1985-2012) which has been processed using a change-detection algorithm called LandTrendr (Kennedy et al., 2010). Outputs from LandTrendr include disturbance patches categorized by agent of change (e.g., clear-cut, partial harvest, etc.) as well as their timing, duration and magnitude. This data will be clipped to the boundary of Willamette National Forest. To structure my analysis, I’ll be using vector data representing the administrative forest boundaries, and the LUDs and ownerships within them.

3. Hypotheses:

My general hypothesis is that patterns of disturbance will vary according to land use designation across space and time. For example, on Matrix Lands, I expect to see spatial clustering of clear-cuts that will increase during the period of the NWFP’s implementation (from 1994 onward), while on Adaptive Management Areas, I expect to see significantly fewer, more dispersed clearcuts, but increases in partial harvests. 

4. Approach:

I will analyze spatial patterns of clear-cut and partial harvest disturbance patches within each LUD (e.g., clustering), as well as the spatial characteristics of these disturbance patches (e.g., edge-to-area ratio). My analysis will be done primarily in ArcGIS, but may also include Python scripting, statistical analysis in R, or “patch analysis” using FRAGSTATS.

5. Expected outcome:

I will produce maps and graphs of disturbance patterns by LUD over time. Hopefully this will help visualize whether or not the NWFP has been implemented as intended. 

 6. Significance:

In the Pacific Northwest, forest governance is shaped largely by the federally mandated Northwest Forest Plan (NWFP), which initiated a momentous shift in forest management priorities; from the provision of sustained timber harvest to protection of ecosystems (Thomas, 2006). The NWFP is implemented through an adaptive management strategy that must identify high priority inventory and monitoring objectives needed to assess the plan’s effectiveness over time (FEMAT, 1993). Ideally, this project and my overall research will contribute to this ongoing assessment.

7. Experience levels with…

ArcGIS = high

Modelbuilder, Python = medium

R = low

References:

Charnley, S., & Pacific Northwest Research Station. (2006). Northwest Forest Plan, the first 10 years (1994-2003) : Socioeconomic monitoring results (General technical report PNW ; 649). Portland, OR: U.S. Dept. of Agriculture, Forest Service, Pacific Northwest Research Station.

Kennedy, R. E., Z. Yang, and W. B. Cohen. 2010. Detecting trends in forest disturbance and recovery using yearly Landsat time series: 1. LandTrendr – Temporal segmentation algorithms. Remote Sensing of Environment 114:2897-2910.

Thomas, J., Franklin, J., Gordon, J., & Johnson, K. (2006). The Northwest Forest Plan: Origins, Components, Implementation Experience, and Suggestions for Change. Conservation Biology, 20(2), 277-287.

Report of the Forest Ecosystem Management Assessment Team (FEMAT, 1993)

1 Research Question

Counties reacted differently towards the Great Recession from Dec. 2007 to June 2009. Economic resilience is defined to measure the performance of counties. This research focused on waht contributes to economic resilience of a county. Especially, how does income inequality affect community economic resilience?

Definition:

Income inequality refers to the uneven manner of income distribution. Gini Coefficient is used to measure income inequality in this analysis. Gini Coefficient measures the ratio of area between lorenz curve and the 45 degree line (perfect equality).

Economic resilience refers to the regional ability to absorb and adjust to an external shock (recession, natural disaster, etc.) Martin (2012) defined dimensions of economic resilience, which two offer guidance for measurement, resistance and recovery(the other two is reorientation and renewal). Resistance means the sensitivity of reaction towards an exogenous shock and recovery shows the speed and degree of recovery.

Rising income inequality might pose an explanation to understand what is behind the great recession and why regions react differently. Economic growth theory shows income inequality is related to economic growth.

• Higher inequality retards growth in poor countries and encourages growth in richer places (Barro, 2000).
• Fallah and Patridge (2007) concludes positive inequality-growth link in the urban sample with the opposite in nonmetro case.
• Inequality leads to lower productivity, more instability, lower efficiency and lower growth (Stiglitz, 2012, Chapter 4).

The assumed relationship is

1. Income inequality(pre-recession, 2000) affects economic resilience.
2. Economic resilience affects income inequality later on (Five year average of 2007-2011 include the recession time). This relationship seems to be weak in rationale. It will not be tested in the analysis.
3. Simultaneous relationship between income inequality and economic resilience. The instrumental variable is hard to find, hence this relationship will not be tested.

The relationships have not been determined yet.

Spatial autocorrelation is assumed to act in the relationship. The effect of income inequality or resilience or demographic or economic factors may not be limited within a region but attenuate with distance. Resilience in a county might be affected by its own characteristics as well as the surrounding counties. (Confused about the difference between GWR, spatial lag and spatial error model.)

1) Identify if counties will be affected by neighboring counties, i.e. spatial clustering of Income Inequality for year 1990, 2000, and ACS 2007-2011, and Economic Resilience/Drop/Rebound.

2) Identify the impact of income inequality on economic resilience or inverse relationship. The simultaneous relationship is hard to test because I haven’t found an IV which affect income inequality but not economic resilience and another which affect economic resilience but not income inequality.

3) How demographic, economic and industrial factors affect income inequality and economic resilience, especially the role of  rural/urban?

4) Spatial Lags or Spatial Errors model

2 Dataset

The cross-sectional data covers 3141 counties in the U.S..  Data for pre-recession time period of  2000 and 2001, are used. Income inequality of year 2000 and ACS 5-Yr Esitmators are used. Details are listed below.

Key variable to measure economic resilience is drop, rebound and resilience, to measure income inequality is Gini coefficient.

Han and Goetz(2015) developed a one number measurement to measure economic resilience ,which was a ratio combined with drop (shows resistance) and rebound(recovery), and called economic resilience. Monthly employment data from Bureau of Labor statistics for 2003-2014 is used to calculate.

http://blogs.oregonstate.edu/geo599spatialstatistics/wp-admin/post.php?post=1605&action=edit

Formulations:

Gini Ceofficient is calculated using Household Income (group means)  from 1990 and 2000 Decennial Census, and American Community Survey 2007-2011 via R using a package inequal. Gini Coefficient provided by American Community Survey for the first time using individual data is used. Income inequality calculate for year 2000 and provided by ACS for 2007-2011 are used in the research.http://blogs.oregonstate.edu/geo599spatialstatistics/wp-admin/post.php?post=1605&action=edit

Key explanatory variable: economic structure(all twenty 3-digit NAICS industry):

Economic Structure : Location Quotients for ten industries chosen by looking at the national level data of annual employment from 2000. They are high in growth

 

3 Hypotheses

1) There is spatial clustering in income inequality and economic resilience.

2) Demographic economic and industrial factors affect income inequality and economic resilience. The relationship differs across rural/urban.

3) Not sure if it is spatial lag or spatial error.

4 Approaches

1) Identify the spatial clusterings of Income Inequality and Economic Resilience: Hot-spot analysis, Global Moran’s I and Anselin Local Moran’s I

2) Identify the impact of income inequality on economic resilience or inverse relationship: OLS, GWR

3) How demographic, economic and industrial factors affect income inequality and economic resilience, especially the role of rural/urban places matter: OLS, GWR

4) Spatial Lags or Spatial Errors model: GeoDa

 

5 Expected outcome

1) Maps of hot spots, clustering of income inequality and economic resilience

2) Statistical relationships between income inequality and economic resilience

3) Statistical relationships between income inequality, economic resilience and other demographic variables.

4) Spatial lag model or spatial error model which fits the statistical relationship of income inequality and economic resilience

 

6 Significance

There are papers discussing income inequality and economic resilience, but little work is done to explore the relationship between income inequality and economic resilience. No spatial analysis so far.

7 Your level of preparation

(a) Arc-Info, medium

(b) Model builder and/or GIS programming in Python, none

(c) R, medium

Reference:

Barro, Robert J. “Inequality and Growth in a Panel of Countries.” Journal of economic growth 5, no. 1 (2000): 5-32.

Fallah, Belal N., and Mark Partridge. “The elusive inequality-economic growth relationship: are there differences between cities and the countryside?.” The Annals of Regional Science 41, no. 2 (2007): 375-400.

Stiglitz, Joseph E. “The price of inequality (London, Allen Lane).” (2012).

Peters, David J. “Income Inequality across Micro and Meso Geographic Scales in the Midwestern United States, 1979–20091.” Rural Sociology 77, no. 2 (2012): 171-202.

 

 

Introduction & Background

Despite the growing number of scientists, federal and state agencies, private citizens, and non-profit organizations working to restore damaged ecosystems in the Great Basin, intact native plant communities continue to decline. The shift away from native-perennial to invasive annual-grass dominated systems has reduced biodiversity, increased wildfire severity and frequency, and has expedited desertification.

To combat this ecosystem overhaul, the most up-to-date and relevant science must be used to guide the restoration of Great Basin plant communities. Improving native plant establishment rates in the restoration setting is one of the biggest challenges faced by land managers. Bluebunch wheatgrass (Pseudoroegneria spicata) is a commonly used native species in restoration but seedling establishment is modest. The goal of our study is to fill knowledge gaps surrounding seedling adaptation to  soils.

The information from a study by St. Clair et al. (2013) was used to delineate seed transfer zones for bluebunch wheatgrass. A seed zone is a geographic area across which native seed can be collected and planted without risking maladaptation. To delineate seed zones, seed is collected from wild populations across the geographic range of a species. Plants are grown to adulthood in several common garden locations that span this geographic range. Climatic data are matched with phenotypes. Each phenotype is defined by observed adult plant traits measured in each common garden (e.g. leaf width, leaf length, pubescence, crown width). Phenotype and climate data are mapped together using a spatial model (Westfall 1992). The final step in creating a seed zone is to empirically verify whether using delineated seed zone maps to guide plant material selection actually enhances plant fitness.

An important knowledge gap I have identified is the role of soils in the adaptation of bluebunch wheatgrass. There exists only a moderate correlation between population-level phenotypic trait variability and seed-source climates, even though traits associated with size, phenology, and leaf morphology varied considerably among the populations sampled (St. Clair et al 2013). Adaptive phenotypic traits are visible expressions of genetic variation caused by environmental factors (i.e. leaf length, flowering phenology, pubescence etc.). This moderate correlation between phenotypic traits and climate may suggest that other factors contribute significantly to local adaptation in bluebunch wheatgrass.

Soils are natural bodies that consist of living organisms, organic matter, minerals, air, water, are comprised of horizons, and have the ability to support plant life (NRCS 2016). The complex array of factors inherent in soils makes generalizing about them challenging. Many studies link plant survival to soil texture class (i.e. percentage sand, silt, and clay) and soil-water dynamics (Letey 1958, Ullah and Hulbert 1969). Others such as Jensen et al. (1990) used soil traits and discriminant analysis to predict sagebrush-dominated plant community types using soil traits such as soil depth, subsoil clay content, total water holding capacity, and A-horizon thickness. Still other studies correlate seedling emergence and germination to aggregate size and bulk density (Nasr and Selles 1995). I aim to determine whether observed phenotypic traits of bluebunch wheatgrass vary with soil traits such as soil texture, soil depth, pH, aggregate structure and water holding capacity.

My study will utilize existing phenotypic trait data and soils maps to explore links between soil order, soil series, plant traits, seed zones, and ecoregions. I predict that phenotypic trait divergence will be correlated to soil gradients that exist across seed zones. Information obtained from this work will either support current seed zones for bluebunch wheatgrass or create better seed zones, and help land managers to achieve higher success rates in restoration.

Description of Datasets

1. The seed zones for bluebunch wheatgrass have been delineated into shapefiles. Each polygon represents a seed zone. The boundaries of these polygons were delineated using ArcMap spatial analysis of climate (precipitation / temperature), elevation, and plant traits (leaf length, leaf width, crown width etc.). This dataset / map was developed in 2013 and is based on two years of data collection at common-garden sites in the Great Basin.
2. The Phenotypic trait dataset contains measurements of individual plants at sixteen different common gardens spread throughout the study area. Each seed zone is represented by two common gardens, and each common garden contains 4-5 populations from each seed zone. Measurements such as leaf length, width, and reproduction stage scores were gathered from each individual plant in all sixteen gardens. Each population of bluebunch wheatgrass growing in the common gardens is associated with a collection site (UTM / elevation / approximate area in acres). Each common garden also has an associated (UTM, elevation / and dimensions in meters).

1. “The gSSURGO (soils) database is derived from the official Soil Survey Geographic (SSURGO) database. SSURGO generally has the most detailed level of soil geographic data developed by the National Cooperative Soil Survey (NCSS) in accordance with NCSS mapping standards. The tabular data represent the soil attributes and are derived from properties and characteristics stored in the National Soil Information System (NASIS). The gSSURGO data were prepared by merging the traditional vector-based SSURGO digital map data and tabular data into statewide extents, adding a statewide gridded map layer derived from the vector layer, and adding a new value-added look up table (valu) containing “ready to map” attributes. The gridded map layer is in an ArcGIS file geodatabase in raster format. The raster and vector map data have a statewide extent. The raster map data have a 10-meter cell size that approximates the vector polygons in an Albers Equal Area projection. Each cell (and polygon) is linked to a map unit identifier called the map unit key. A unique map unit key is used to link the raster cells and polygons to attribute tables” (Soil survey staff 2015).

Hypotheses

I hypothesize that some phenotypic trait variability in bluebunch wheatgrass can be explained by adaptation to certain soil traits. My objective is to determine if phenotypic trait variability in bluebunch wheatgrass corresponds to soil order, soil series, or other distinct soil trait groupings, and if the existing seed zones for bluebunch wheatgrass and/or ecoregion account for the spatial distribution of these soil trait groups.

Approaches

1. Soil maps containing soil order polygons will be compared to existing bluebunch wheatgrass seed zone polygons to determine the extent to which soil order is contained within existing seed zones.
2. Plant trait data from all 16 common gardens will be compared to the soil traits (i.e. order, series, texture class or other) to look for correlations between source population soils and genetic expressions in bluebunch wheatgrass.
3. Soil maps containing soil series polygons will be compared to existing bluebunch wheatgrass seed zone polygons.
4. Available soil maps containing soil order and series will be compiled for our study area. Soil series and pedon descriptions will be used to assemble a set of soil traits that are relevant to plant growth and that exist in the study area. These traits will include, depth to bedrock, A and B horizon texture classes, percent gravel, pH, aggregate structure, and depth to impermeable layers. These data will be organized into a database for statistical analysis in PC-Ord (a multivariate statistical analysis software). Phenotypic trait data (such as leaf length, basal width, and seed production) will be gathered from a previous study involving sixty bluebunch wheatgrass populations grown in sixteen common gardens spread across the study area (figure 1).

Expected Outcomes

I expect to find that bluebunch wheatgrass phenotypes are correlated with soil order and may be correlated with other soil traits such as texture and aggregate structure.

Significance

A relationship between soil and bluebunch wheatgrass phenotypic expression suggests that soils information should factor into the seed-zone delineation process.

Level of preparation

I have moderate experience with ArcMap but have never used the program to do spatial analysis. Similarly, I have two terms of introductory statistics that used R but have never explored multivariate datasets.

References:

Predeville, Holly. 2016. Depiction of data gathered in a study by St. Clair et al. 2013 in google maps. https://www.google.com/maps/d/edit?mid=zxFzk9yulKc0.klEzA-cxckLY

Jensen, M. E., G. H. Simonson, and M. Dosskey. 1990. “Correlation between Soils and Sagebrush-Dominated Plant Communities of Northeastern Nevada.” Soil Science Society of America Journal 54 (3): 902. doi:10.2136/sssaj1990.03615995005400030049x.

Letey, J. 1958. “Relationship between Soil Physical Properties and Crop Production.” In Advances in Soil Science, edited by B. A. Stewart, 277–94. Advances in Soil Science 1. Springer New York. http://link.springer.com.ezproxy.proxy.library.oregonstate.edu/chapter/10.1007/978-1- 4612-5046-3_8.

Nasr, H. M., and F. Selles. 1995. “Seedling Emergence as Influenced by Aggregate Size, Bulk Density, and Penetration Resistance of the Seedbed.” Soil and Tillage Research 34 (1): 61–76. doi:10.1016/0167-1987(94)00451- J.

NRCS 2016. What Is Soil? | NRCS Soils. 2016. Accessed March 3. http://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/edu/?cid=nrcs142p2_054280.

Soil Survey Staff. Gridded Soil Survey Geographic (gSSURGO) Database for the Conterminous United States. United States Department of Agriculture, Natural Resources Conservation Service. Available online at https://gdg.sc.egov.usda.gov/. November 16, 2015 (FY2016 official release).

St. Clair Bradley John, Francis F. Kilkenny, Richard C. Johnson, Nancy L. Shaw, and George Weaver. 2013. “Genetic Variation in Adaptive Traits and Seed Transfer Zones for Pseudoroegneria Spicata (bluebunch Wheatgrass) in the Northwestern United States.” Evolutionary Applications 6 (6): 933–48. doi:10.1111/eva.12077.

Ullah Alizai Hamid, and Lloyd Hulbert. 1969. “Effects of soil texture on evaporative loss and available water”, Soil Science.

Westfall, R. D. 1992. “Developing Seed Transfer Zones.” In Handbook of Quantitative Forest Genetics, edited by Lauren Fins, Sharon T. Friedman, and Janet V. Brotschol, 313–98. Forestry Sciences 39. Springer Netherlands. http://link.springer.com.ezproxy.proxy.library.oregonstate.edu/chapter/10.1007/978-94-015-7987-2_9.