Biol 515 Landscape Ecology and Management



Biol 515 Landscape Ecology and Management, Fall 2011

Lab #5

QUANTIFYING SPATIAL PATTERN II

Objectives:

1. Gain experience in selecting appropriate landscape metrics and using them to answer ecological questions.

2. Reflect on the relative effects of human disturbance and natural disturbance in influencing landscape patterns.

Overview of Lab 5:

In this lab, you will answer a question about landscape pattern in a real-world landscape and discuss the ecological implications of the patterns using landscape metrics.

Homework: The write up from Part II is due by the beginning of the next lab, Oct 10.

Introduction

Natural versus Human Disturbance

Many ecological systems require natural disturbance like wildlife or flooding to maintain successional stages and ecological function. This is likely true because disturbance regimes have been in place long enough to be selective forces and organisms have adapted to these disturbance regimes.

Human activities such as logging are often thought to disrupt natural landscapes and decrease habitat suitability of native species. Yet these human disturbances share many features with natural disturbance and thus may have similar effects on ecological function. They initiate succession and ultimately result in a range of seral stages and complex landscape patterns. However, some ecologists suggest that human disturbance often differs from natural disturbance in severity, patch size and shape, and in return interval. Thus, it pushes the landscape to a very different condition than natural disturbance.

These questions suggest the need to compare human and natural disturbance, evaluate its effects on landscape patterns, and to quantify consequences for biodiversity and ecological function. The resulting knowledge would provide a basis for managing human and natural disturbances to best achieve ecological objectives.

Study Area

The western boundary of Yellowstone National Park (YNP) is striking as a dividing line between a landscape largely driven by wildfire and one driven by clearcut logging. Within YNP, vast forests of primarily lodgepole pine and subalpine fir vary in seral stage due to wildfire. The natural fire regime is comprised of relatively severe crown fires that burn large areas with a return interval of 200-300 years. The most recent of these large fires was in 1988 when some 43% of YNP burned. The current YNP landscape is a mosaic of old-growth forests and early seral patches now recovering from the 1988 fire.

[pic]

Figure 1. Map of the western portion of YNP and Targhee National Forest. The red in A and B show: A - wildfire in forest, B – clearcut areas in forest. Undisturbed areas in YNP are shown in C. The general area of our focus for this analysis is shown by the blue box.

The adjacent Targhee National Forest supports a similar forest type to YNP.

Management here has focuses on excluding wildfire and promoting timber harvest. The timber harvest was motivated both by the objective of producing wood products for local mills and by interest in removing the large number of trees that were killed by a bark beetle outbreak in the 1970s. The harvest system was clearcutting in a staggered setting. By spacing the harvest units across the landscape, negative impacts on say erosion are thought to be minimized and beneficial effects of providing edge habitats for some game species are maximized.

Wildlife species differ in habitat requirements. Elk, for example, use both older forest for cover and recently disturbed areas for foraging. The American marten requires mature and old growth forest with high levels of coarse woody debris and multiple canopy layers. They avoid the edges of forest patches that are adjacent to clearings because of altered microclimate. The black-backed woodpecker is an obligate on recently burned forest.

Research Question and Exercise

The contrasting YNP and TNF landscapes offer a unique opportunity to quantify the relative effects of natural and human disturbance on landscape patterns and ecological function. Ecologists have suggested that wildfire produces landscapes that are more heterogeneous (patch size, patch shape), less fragmented, and better connected than staggered setting logging. This leads to the following hypotheses:

Compared with staggered setting logging, wildfire results in landscapes are where:

1. disturbance patches are more variable in patch size and more complex in shape;

2. forest patches are less fragmented (total area of patch type, core area, edge to interior ratio);

3. late seral habitats are better connected (nearest neighbor distance).

Test these hypotheses for the study area and interpret the results for the three species of wildlife described above. Select two landscape metrics for each of the three hypotheses that you think are best suited to testing the hypotheses. Run V-LATE for each landscape to quantify these metrics. Plot the results and incorporate them into a short report that addresses the hypotheses and consequences for wildlife.

Exercise

**Note, specific tasks are not given great detail since we used V-LATE in last weeks lab. Feel free to use previous labs for specific instructions.

Part I: Inspecting Landscape Patterns and Selecting Landscape Metrics

Copy the folder ‘lab5_2011’ from a flash drive (provided in class) to C:\Users\bio515.

Look (in ArcGIS) at the patterns of forest and disturbance patches in the TNF and YNP landscapes and qualitatively consider how they are similar or different relative to the hypotheses and the wildlife species.

1) Add the boundary file for the Yellowstone National Park to the MDA (ynpbdyvector.shp). Change the polygon fill to ‘no fill’ (double click, Layer Properties, Symbology) and Zoom to this layer to understand where you will be analyzing later.

2) Now add the landscape files. Ynpvector.shp and tnfvector.shp. Zoom in to these layers so you can see both grids at the same time.

The two polygon landscape files are two landscapes the polygons that represent landcover type. Since the two grids currently have ‘single symbol’ legends we will change the legend type to “Categories” so that each landcover type is represented by a color of your choice (Layer Properties, Show: Unique Values, Value Field: Gridcode; Add All Values). Modify the “Symbol” colors and “Labels” for both files for comparison across the two landscapes (i.e modify the legend so that both landscapes have the same color scheme and labels for classes below. Since the ‘other’ land cover class is not of interest, make this class white or hollow.

For the TNF landscape the legend is:

Value 1 = other land cover class

Value 2 = Forest

Value 3 = Disturbed_logged

For the YNP landscape the legend is:

Value 1 = other land cover class

Value 2 = Forest

Value 3 = Disturbed_burned

3) Save your ArcGIS project (File, Save As).

Based on your qualitative evaluation and the descriptions of landscape metrics from last week’s lab (Available on the class webpage if needed), select 2 landscape metrics from Table 1 below, that you think would best test each of the hypotheses above. Use V-LATE as we did in class to calculate the metrics. A description of all possible V-LATE metrics can be found in Appendix 1.

For each hypothesis, choose one metric that provides the best test. Import these into Microsoft Excel and produce graphs or charts. You may wish to create an Excel file that contains output for both the TNF and YNP landscapes by cutting and pasting. This will allow for plots of side by side comparisons of the metrics for the two landscapes.

Table 1. Below are examples of metrics to test the above hypotheses.

|Metric |Code |Patch Class |Landscape |

| | | |TNF |YNP |

|Percent of landscape |PLAND |Disturbed | | |

| | |Forest | | |

|Mean patch area |AREA_MN |Disturbed | | |

| | |Forest | | |

|Standard deviation in |AREA_SD |Disturbed | | |

|patch size | | | | |

| | |Forest | | |

|Shape index – mean |SHAPE_MN |Disturbed | | |

| | |Forest | | |

|Total core area (assume |TCA |Disturbed | | |

|100 m edge depth) | | | | |

| | |Forest | | |

|**Core area percent of |CPLAND |Disturbed | | |

|landscape | | | | |

| | |Forest | | |

|**Core area – mean |CORE_MN |Disturbed | | |

| | |Forest | | |

|Edge density |ED |Disturbed | | |

| | |Forest | | |

|Euclidean nearest |ENN_MN |Disturbed | | |

|neighbor distance - mean | | | | |

| | |Forest | | |

|Aggregation index |AI |Disturbed | | |

| | |Forest | | |

**specify the ‘edge’ distance for the core area metrics as 100

Part III: Interpreting Results

Write a short report (Text: no more than 1 single spaced page, 12 pt font, 1 in margins) that address the hypotheses and consequences for the wildlife species. Include supporting graphics. Sections should include: statement of problem; methods, results, discussion, graphics.

Appendix I.

Description of V-LATE Calculationws

1. Calculates Area and Perimeter for the selected polygon shapefile. The adumbrated tree structure of the calculations [1] and [4] - [9] provides a guideline for a logical order of analysis. If you, for example, want to perform Edge Analysis you'll have to calculate Area/Perimeter first. For any of the categories [4] - [9] you'll need to calculate Area/Perimeter only one time in the beginning.

2. Proximity Analysis: Nearest Neighbour calculation for selected classes. Three fields are added to the attribute table: Distance to Nearest Neighbour (NNDist), the ID of the NN polygon (NNID) and the Area of the NN (NN_Area). Proximity (fragstats) calculates proximity as implemented in Fragstats (McGarigal & Marks, 1995). A respective field is added to the attribute table (PXfg_PB - where PB stands for the chosen Proximity Buffer size). Results on class/landscape level can be exported to a text file (*.txt).

3. Class Info: Shows the currently chosen class field and selected classes. Use the change buttons to change class field or classes.

4. Area Analysis: Calculates Number of patches (NP), Class Area in m² (CA), Mean Patch Size in m² (MPS), Patch Size Standard-Deviation in m² (PSSD) for selected classes. Results can be exported to a text file (*.txt).

5. Edge Analysis: Calculates Mean Patch Edge in m (MPE), Total Edge in m (TE) and Edge Density in m/ha (ED) for selected classes. Results can be exported to a text file (*.txt).

6. Form Analysis: Calculates Mean Shape Index (MSI), Mean Perimeter-Area ratio (MPAR), Mean Fractal Dimension (MFRACT) on class/landscape level and Fractal Dimension (FD), Perimeter-Area Ratio (PARATIO), Shape Index (SHAPE) on patch level for selected classes. Results can be saved as textfile (*.txt). The results on patch level are written into three different fields to the attribute table (Paratio, Shape_Idx and Frac_Dim). If the Patch Area is equal or smaller than 1 [map units], the Frac_Dim value will be set to -999 ("No Data").

7. Core Area Analysis: Creates a new theme (layer) that represents the interior of a patch (core area) based on a user-specified buffer distance. It will produce an output table. Calculates Total Area in m² (TA), Total Core Area in m² (TCA), Core Area Index in % (CAI), Number of Core Areas (NCA), Total Class Core Area in m² (TCCA) and CORITY (equals the # of patches minus # of patches without core area divided by # of resulting core area (including size = 0!)). Results can be exported to a text file (*.txt).

Note that the resulting core area may differ from core area calculation in ArcView 3.x due to differing algorithms. It seems that buffering in ArcGIS is now more accurate.

8. Diversity Analysis: Calculates Richness, Shannon´s Diversity Index, Shannon´s Evenness Index, Dominance and Proportion of class for selected classes. Results can be exported to a text file (*.txt).

9. Subdivision Analysis: Calculates Landscape Division Index (DIVISION), Splitting Index (SPLIT) and Effective Mesh Size (MESH). Results can be exported to a text file (*.txt).

10. Dissolve ID: Creates a Dissolve ID for adjacent polygons of the same class. Based on the Dissolve_ID one can dissolve adjacent polygons of the same class using the Dissolve Tool of ArcGIS. Note that since ArcGIS version 9 the user can choose between a Dissolve with multipart and without multipart polygons. Multipart shapefiles though are not suitable for V-LATE calculations.

11. Delete Fields: Delete one or several fields in the attribute table.

12. IDEFIX Database: Launches the indicator database IDEFIX database, if locally available under the specified path.

13. Help: Opens this help file (if stored in the same folder as the V_LATE_2beta_argis10.dll file)

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download