You have been hired by a conservation group to determine how selective logging practices have changed a rainforested area in the Congo Basin of Central Africa. You must use ArcGIS and Spatial Analyst to determine the number of forest fragments that have been created by logging roads. You also need to characterize the habitat quality of each forest fragment in terms of the ratio of interior to edge habitat, the edge to area ratio, the thickness, and the overall area.
At the successful completion of Lesson 7, you will have:
If you have questions now or at any point during this lesson, please feel free to post them to the Lesson 7 Discussion.
This lesson is one week in length and is worth a total of 100 points. Please refer to the Course Calendar for specific time frames and due dates. To finish this lesson, you must complete the activities listed below. You may find it useful to print this page out first so that you can follow along with the directions. Simply click the arrow to navigate through the lesson and complete the activities in the order that they are displayed.
SDG image retrieved from the United Nations [1]
Tropical rainforests are extremely valuable in terms of the ecological services they provide, such as biodiversity and carbon sequestration. Although they cover only 6% of the earth's surface, they provide habitat for over half of the plants and animals in the world. Many of these plants and animals are threatened or endangered species. Rainforests are also highly prized for their commercial hardwood trees. The trees are cut down and processed to create products such as teak and mahogany furniture, plywood, and flooring.
Historically, there are two main types of logging practices used in tropical forests: clearcutting, and selective logging. During clearcutting, loggers remove all of the trees in a given area, leaving large clearings in their wake. These clearings reduce the amount of usable habitat for plants and animals. You can view an example of clearcutting in Brazil by looking in Google Maps [2]. Notice the large grey patches in the images. These are areas of the forests that were cleared of all vegetation. Without vegetation to stabilize the soil, winds, and rain quickly erode the nutrient-rich soils required for new species to colonize the area. Clearcutting is a very environmentally destructive process. While on Google Maps, be sure to zoom out a little, and you will see the fishbone pattern that is characteristic of logging in tropical forests. Also, check out the article "Roads could help protect the environment rather than destroy it, argues Nature Paper [3]."
In contrast, only species of value are extracted from the forest during selective logging. It seems like this process would be much more environmentally friendly, considering that much of the logged area remains forested. However, it is also a destructive process because it opens up previously inaccessible areas to human exploitation, damage, and degradation. Once logging roads are built, they tend to be used for many other activities. For example, bushmeat hunters use new roads to extract and transport illegal forest products such as monkeys, gorillas, and chimpanzees. Migrating people often travel along these roads and establish new villages. Once settled, they tend to clear surrounding areas of the forest for agriculture using slash-and-burn techniques.
As loggers build new roads, they break up large tracts of forest into progressively smaller areas or “patches.” This process is known as "forest fragmentation." Scientists use a quantity called the "edge to area ratio" to characterize forest fragments. The measurement, calculated as the perimeter of forest/area of forest, represents the complexity of the shape of each forest patch. The higher the value, the more irregular the forest boundary.
In addition to breaking up forests into smaller patches, road building activities also increase disturbances known as "edge effects." Some examples of edge effects include changes in species composition, diversity, and seed dispersion, increased tree mortality and susceptibility to fires, microclimate shifts (humidity and sunlight), increased carbon emissions, and impeding movement of animals. Scientists have observed edge effects up to 2 km from road edges.
Logging activities can have a significant impact on the local ecosystem since the smaller forest patches do not provide the same quantity and quality of habitat as large tracts. As new roads are built, fragments of forests are further degraded as the ratio of interior habitat to edge habitat decreases. Native animal species of tropical rainforests can require blocks of interior habitat greater than 1,000 sq km. Large mammals and species under hunting pressure can require interior areas of at least 10,000 sq km. To create large areas of interior habitat, care must be taken to limit road building activities to certain areas.
In Lesson 7, we are going to examine the effects of historical commercial logging activities for a forested area in southeastern Cameroon. The study area is part of the Congo Basin, which contains the world's second-largest concentration of tropical rainforests. The study area boundary encompasses two main types of land management areas: protected areas and logging areas. The Lobeke National Park is the main protected area within the study site. Using GIS datasets showing road centerlines, we will use ArcGIS Spatial Analyst to quantify the fragmentation and edge effects within the study area.
There are three types of required readings for Lesson 7: a short video describing issues facing forests and the conservation group - World Resources Institute (WRI) that created the data we will use in the lesson, background information about the study area, and Esri Help Topics related to the GIS tools we will use in the lesson.
Note: If you do not see the embedded video below, try clicking the refresh button on your browser, try viewing this page in another browser, or click on the hyperlink listed as the source.
Find the help articles listed below in the ArcGIS Pro Resources Center [5].
Search for:
This section provides links to download the Lesson 7 data along with reference information about each dataset (metadata). Briefly review the information below so you have a general idea of the data we will use in this lesson.
Note: You should not complete this activity until you have read through all of the pages in Lesson 7. See the Lesson 7 Checklist for further information.
Create a new folder in your GEOG487 folder called "Lesson7." Download a zip file of the Lesson 7 Data [15] and save it in your "Lesson7" folder. Extract the zip file and view the contents. Information about all datasets used in the lesson is provided below:
The data for this lesson is contained in a geodatabase called Lesson7.gdb. Read about geodatabases in the online help provided by Esri if you are not familiar with this data format.
The roads and study boundary were sourced from portions of the Interactive Forestry Atlas [16], Global Forest Watch Open Data Portal [17], Humanitarian Data Exchange [18], and OpenStreetMap [19] for our study area. The original data was created by the World Resources Institute - Global Forest Watch [20] (GWF), a nongovernmental organization (NGO) focused on environmental issues. They regularly produce reports and data about the state of forests and logging in Central Africa and other locations around the world. Part of their work involves the creation of GIS datasets to assist forest managers.
In Part I, we will review the historical data and organize the map for analysis. In Part II, we will use the roads dataset to create rasters of habitat quality and forest patches. In Part III, we will generate statistics about the size, shape, and habitat quality of each forest patch. We will also generate statistics of habitat quality by land management type (conservation vs. logging areas). In Part IV, we will share our analysis results using ArcGIS Online.
Note: You should not complete this step until you have read through all of the pages under the Lesson 7 Module. See the Lesson 7 Checklist for further information.
In Part I, we will review the data and organize the map for analysis.
Since all of the datasets used in this lesson have the same projection, we do not have to be concerned with the order in which we load the data.
How many of the management units are used for logging? What about conservation?
Using the "Imagery Hybrid" layer, can you see the approximate extent of the rainforests located in the Congo Basin? What kind of details can you see in the forest if you zoom in very close?
In Part II, we will use the roads dataset to create raster data layers of habitat quality and forest patches. In Part III, we will generate statistics about the size, shape, and habitat quality of each forest patch.
We will use the following coded values:
When you convert a feature layer to a raster, you have to choose a field in the feature layer from which to base the grid cell values on. You often need to create a new dummy field and assign a value that is consistent for all of the records you want to convert (like we did above).
It is also important to note that if there are any selected records in the vector layer, only those records will be converted to a raster layer. Therefore, be sure to clear any selected features before performing the conversion.
The data type of the field you choose is very important. For example, if you choose a numerical field that contains decimal values, the resultant grid will not have an attribute table. However, if you choose an integer field, the resultant raster will have an attribute table. If you choose a text field, ArcGIS will automatically assign each unique text value an integer code in a new field named "VALUE."
The new raster layer will be created based on all defined Spatial Analyst environment settings. Always check these settings before converting features to a raster to avoid potentially undesirable results.
It is important to note that although the extent setting is utilized by Feature to Raster, the mask setting is ignored. Although you will not notice this with the "RoadsGrid.tif" layer, you will see the effects of this when you create a buffered grid later in this lesson.
Make sure you have the correct answer before moving on to the next step.
The "RoadsGrid.tif" raster should have the following information. If your data does not match this, go back and redo the previous step.
Remember from the Background Information section that edge effects can occur up to 2 km from roads. We will consider all areas 2 km from roads as "edge habitat" and areas farther than 2 km from roads as "interior habitat." To do this, we need to create a buffer of the road centerlines.
Make sure you have the correct answer before moving on to the next step.
The "EdgeGrid" raster should have the following information. If your data does not match this, go back and redo the previous step.
Did the Reclassify Tool honor the mask and extent settings?
Hint: Compare the InteriorGrid.tif and EdgeGrid.tif rasters along the study area boundary.
Make sure you have the correct answer before moving on to the next step.
The "InteriorGrid.tif" grid should have the following information. If your data does not match this, go back and redo the previous step.
In steps 1, 2, and 3, we created three individual grids, one for each level of habitat quality. To continue the analysis, we need a way to merge all of the data sets into one grid. The Mosaic to New Raster tool in Toolboxes will allow you to mosaic multiple raster data layers together by stacking them on top of one another. The values in the output raster will be determined based on the order the files are specified during the mosaic. Cells will first be assigned according to the cell values in the first input raster; all remaining null values will be filled in with the middle input raster, and so on. We want the roads to be on top of the stack, the edge habitat in the middle, and the forests on the bottom.
This tool does not honor the Output extent environment settings. If you want a specific extent for your output raster, consider using the Clip tool. You can either clip the input rasters prior to using this tool, or clip the output of this tool.
Make sure you have the correct answer before moving on to the next step.
The "HabMosaic" raster should have the following information. If your data does not match this, go back and redo the previous step.
What value was assigned to areas with roads, since they have data in both the "RoadsGrid" and "EdgeGrid" rasters?
Which habitat type (roads, edge, or interior) covers the majority of the study area?
How can you calculate the area of each habitat type?
The Raster Calculator utilizes all raster environment settings, so it is highly useful when working with raster data. As displayed above, simply selecting a raster layer and running the Raster Calculator will generate a new raster layer based on the current environmental settings. Try changing these settings to see the differences when running the Raster Calculator on a particular raster layer.
Make sure you have the correct answer before moving on to the next step.
The "HabitatGrid" raster should have the following information. If your data does not match this, go back and redo the previous step.
We now have one grid with values showing the range of habitat quality within the study area. The next step is to create a grid of forested areas, which we need to create the forest fragments. We will use the "RoadsGrid.tif" raster we created in Part II Step 1 to create a new grid representing forested areas (cells that are NOT roads).
Make sure you have the correct answer before moving on to the next step.
The "ForestGrid.tif" raster should have the following information. If your data does not match this, go back and redo the previous step. You may need to adjust for the Mask and Processing Extent here as well.
Make sure you have the correct answer before moving on to the next step.
The "ForestPatches.tif" grid should have the following information. If your data does not match this, go back and redo the previous step.
OID | Value | Count | link |
---|---|---|---|
0 | 1 | 64201 | 1 |
1 | 2 | 58 | 1 |
2 | 3 | 122867 | 1 |
3 | 4 | 19 | 1 |
4 | 5 | 30 | 1 |
Why did we use the number "100" to calculate the area?
Make sure you have the correct answer before moving on to the next step.
The "ForestPatches.tif" grid should have the following information. If your data does not match this, go back and redo the previous step.
oid | value | count | forestid | area_sq |
---|---|---|---|---|
0 | 1 | 64201 | 1 | 642010000 |
1 | 2 | 58 | 2 | 580000 |
2 | 3 | 122867 | 3 | 1228670000 |
3 | 4 | 19 | 4 | 190000 |
4 | 5 | 30 | 5 | 300000 |
5 | 6 | 1 | 6 | 10000 |
6 | 7 | 13 | 7 | 130000 |
7 | 8 | 1 | 8 | 10000 |
8 | 9 | 318427 | 9 | 3184270000 |
How many individual forest patches are there? Which forest patch is the largest? Which forest patch is the smallest? Why do you think there are so many patches with an area of exactly 10,000 sq m?
In Part III, we will use two Spatial Analyst tools to bring together the raster layers we created in Part I (habitat quality) and Part II (forest patches). Zonal Geometry calculates several geometry measures, such as area and thickness, for zones in a raster. We will use it to generate a table of statistics about the size and shape of each forest patch. We will also use the Zonal Histogram Tool to tabulate the number of cells of each habitat type within each forest patch and management unit.
Make sure you have the correct answer before moving on to the next step.
The "PatchGeometry" table should have the following information. If your data does not match this, go back and redo the previous step.
OID | Value | Area | Perimeter | Thickness | Xcentroid | ycentroid | Majoraxis | minoraxis | orientation |
---|---|---|---|---|---|---|---|---|---|
0 | 1 | 642010000 | 350600 | 6343.9 | 1688100 | 359206 | 24878.6 | 8214.21 | 81.9311 |
1 | 2 | 580000 | 3800 | 212.1 | 1696040 | 379519 | 631.205 | 292.488 | 84.2196 |
2 | 3 | 1228670000 | 907000 | 6250.3 | 1756350 | 335775 | 36173.2 | 10811.8 | 134.675 |
3 | 4 | 190000 | 2400 | 150 | 1698610 | 378516 | 270.871 | 223.275 | 140.531 |
4 | 5 | 300000 | 2800 | 170.7 | 1699130 | 378353 | 401.382 | 237.911 | 110.363 |
5 | 6 | 10000 | 400 | 50 | 1699560 | 378016 | 56.419 | 56.419 | 90 |
6 | 7 | 130000 | 1600 | 150 | 1700800 | 377131 | 219.193 | 188.785 | 166.224 |
7 | 8 | 10000 | 400 | 50 | 1698360 | 377216 | 56.419 | 56.419 | 90 |
Which field in the "PatchGeometry" table is the equivalent to the "ForestID" field? What are the units of the fields "AREA," "PERIMETER," and "THICKNESS"? What do the values in the fields "XCENTROID," "YCENTROID," "MAJORAXIS," "MINORAXIS", and "ORIENTATION" mean?
The Zonal Histogram tool will create a summary table that contains one row for each unique value in the "Value raster" and one column for each unique value in the "Zone dataset." The tool will calculate the total number of cells for each combination of a unique row and column. The tool can also create a graph based on the output table, which we are going to skip.
Make sure you have the correct answer before moving on to the next step.
The "Habitat_by_Patch" table should have the following information. If your data does not match this, go back and redo the previous step.
oid | Label | Value_2 | Value_3 | FORESTID | EDge_sqm | int_sqM | PCTtotedge | pcttotint |
---|---|---|---|---|---|---|---|---|
0 | 1 | 22207 | 41994 | 1 | 222070000 | 419940000 | 35 | 65 |
1 | 2 | 58 | 0 | 2 | 580000 | 0 | 100 | 0 |
2 | 3 | 74095 | 48772 | 3 | 740950000 | 487720000 | 60 | 40 |
3 | 4 | 19 | 0 | 4 | 190000 | 0 | 100 | 0 |
4 | 5 | 30 | 0 | 5 | 300000 | 0 | 100 | 0 |
What do numbers in the "LABEL" field of the "Habitat_by_MU" mean? Which management unit "use" has the most roads?
Make sure you have the correct answer before moving on to the next step.
The "Habitat_by_MU" table should have the following values. If your data does not match this, go back and redo the previous step.
OID | Label | logging | coservation | habitat | logSqm | conssqm | pcttotlog | pcttotcons |
---|---|---|---|---|---|---|---|---|
0 | 1 | 35322 | 1428 | Low Quality Habitat | 353220000 | 14280000 | 96 | 4 |
1 | 2 | 635978 | 44954 | Medium Quality Habitat | 6359780000 | 449540000 | 93 | 7 |
2 | 3 | 302425 | 167611 | High Quality Habitat | 3024250000 | 1676110000 | 64 | 36 |
Make sure you have the correct answer before moving on to the next step.
The "forestpatchpoly" shapefile should have the following information. If your data does not match this, go back and redo the previous step. Note that this table has been sorted based on "gridcode".
Fid | Shape* | Id | gridcode | ForestID |
---|---|---|---|---|
36 | Polygon | 37 | 1 | 1 |
0 | Polygon | 1 | 2 | 2 |
61 | Polygon | 62 | 3 | 3 |
1 | Polygon | 2 | 4 | 4 |
2 | Polygon | 3 | 5 | 5 |
Why is there such a large range of values for the edge to area ratio results?
How would the results of the analysis change if we used a larger or smaller cell size?
Make sure you have the correct answer before moving on to the next step.
The "Final_Forest_Patches" attribute table should have the following information. If your data does not match this, go back and redo the previous step.
FID | Shape* | Forest ID | Totareasqm | perimeterm | thichnessm | edge_sqm | int_sqm | pcttotedge | pcttotint | edgetoarea |
---|---|---|---|---|---|---|---|---|---|---|
36 | Polygon | 1 | 642010000 | 350600 | 6343.9 | 222070000 | 419940000 | 35 | 65 | 0.05461 |
0 | Polygon | 2 | 580000 | 3800 | 212.1 | 580000 | 0 | 100 | 0 | 0.655172 |
61 | Polygon | 3 | 1228670000 | 90700 | 6250.3 | 740950000 | 487720000 | 60 | 40 | 0.07382 |
1 | Polygon | 4 | 190000 | 2400 | 150 | 190000 | 0 | 100 | 0 | 1.26316 |
2 | Polygon | 5 | 300000 | 2800 | 170.7 | 300000 | 0 | 100 | 0 | 0.933333 |
3 | Polygon | 6 | 10000 | 400 | 50 | 10000 | 0 | 100 | 0 | 4 |
5 | Polygon | 7 | 130000 | 1600 | 150 | 130000 | 0 | 100 | 0 | 1.23077 |
Notice how the default outputs from many of the Spatial Analyst tools are not very easy to understand. It’s worth the time to create more intuitive fields, units, and names while you are doing the analysis. That way you can easily interpret your results later on and share them with others in a meaningful format.
In Part IV, we will finalize our map in ArcGIS, then you will be asked to share your results with the Geog487 AGO group as web maps. As a final step, you will combine the output from the Step-by-Step and Advanced Activity into a web application.
That’s it for the required portion of the Lesson 7 Step-by-Step Activity. Please consult the Lesson Checklist for instructions on what to do next.
Try one or more of the optional activities listed below.
Landsat satellite images were used to digitize the road data we used in this lesson. You can read more about Landsat data on NASA’s website [23]. As of October 2008, Landsat data is available for free to the public. It can be viewed and downloaded from the USGS Earth Explorer Viewer [24].
Note: Try This! Activities are voluntary and are not graded, though I encourage you to complete the activity and share comments about your experience on the lesson discussion board.
Advanced Activities are designed to make you apply the skills you learned in this lesson to solve an environmental problem and/or explore additional resources related to lesson topics.
In the Step-by-Step activity, we used road center-lines from the year 2007 to explore forest fragmentation and edge effects. Using the road centerlines from 2001 (roads01), how many forest patches were located in the study site in 2001? Using the road centerlines from 2021 (roads21), how many forest patches were located in the study site in 2021?
In Lesson 7, we used Buffers, the Reclassify Tool, RegionGroup, ZonalGeometry, and Zonal Histograms to explore how logging roads have degraded tropical rainforests in southeastern Cameroon. Specifically, we determined how many forest patches were created, the area and shape of each forest patch, their edge/area ratio, and the area of edge and interior habitat. We also summarized the habitat type by management unit to see whether conservation areas or logging areas provide the best habitat.
Lesson 7 is worth a total of 100 points.
Peer Review (optional): Explore other students' submission and add a short comment on their discussion post.
2007 Forest Patches Map | Web map is posted to ArcGIS Online and link is made available in Canvas. Map is sufficiently designed and described. (20pts) | Map is linked, but it is missing an element or two (layers, descriptions, symbology, etc.) (15pts) | Map is linked but is missing several elements (map, layers, description) or is poorly designed. (10pts) | Link is missing. (0pts) | 20pts |
---|---|---|---|---|---|
2001 Forest Patches Map | Web map is posted to ArcGIS Online and link is made available in Canvas. Map is sufficiently designed and described. (20pts) | Map is linked, but it is missing an element or two (layers, descriptions, symbology, etc.) (15pts) | Map is linked but is missing several elements (map, layers, description) or is poorly designed. (10pts) | Link is missing. (0pts) | 20pts |
2021 Forest Patches Map | Web map is posted to ArcGIS Online and link is made available in Canvas. Map is sufficiently designed and described. (20pts) | Map is linked, but it is missing an element or two (layers, descriptions, symbology, etc.) (15pts) | Map is linked but is missing several elements (map, layers, description) or is poorly designed. (10pts) | Link is missing. (0pts) | 20pts |
Web Map Application | Web app link is posted and made available in Canvas. App facilitates the direct comparison between at least two of the maps (e.g., the 2007 and 2001 maps or the 2007 and 2021 maps). Symbology is consistent between the two maps. Description of the app and maps sufficiently orients the reader and helps to convey trends. (20pts) | Web app link is posted. Some elements of the assignment are missing, but the app still allows for map comparison. (15pts) | Web app link is present but is missing several elements, does not function properly, or otherwise impairs the ability to compare the two maps. (10pts) | Web app link is missing. (0pts) | 20pts |
TOTAL | 80pts |
If you have anything you'd like to comment on, or add to the lesson materials, feel free to post your thoughts in the Lesson 7 Discussion. For example, what did you have the most trouble with in this lesson? Was there anything useful here that you'd like to try in your own workplace?
This page includes links to resources such as additional readings and websites related to the lesson concepts. Feel free to explore these on your own. If you'd like to suggest other resources for this list, contact the instructor
Links
[1] https://www.un.org/sustainabledevelopment/news/communications-material/#FAQ
[2] http://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=brazil+clear+cut&sll=-3.873161,-54.241219&sspn=0.135817,0.125828&ie=UTF8&filter=0&rq=1&ev=zi&radius=4.34&ll=-3.875216,-54.241219&spn=0.140098,0.125828&t=h&z=13
[3] https://news.mongabay.com/2013/03/roads-could-help-protect-the-environment-rather-than-destroy-it-argues-nature-paper/
[4] https://onlinelibrary.wiley.com/doi/full/10.1002/ece3.7027
[5] https://www.esri.com/en-us/arcgis/products/arcgis-pro/resources
[6] https://pro.arcgis.com/en/pro-app/tool-reference/conversion/converting-features-to-raster-data.htm
[7] https://pro.arcgis.com/en/pro-app/tool-reference/conversion/an-overview-of-the-from-raster-toolset.htm
[8] https://pro.arcgis.com/en/pro-app/tool-reference/conversion/an-overview-of-the-to-raster-toolset.htm
[9] https://pro.arcgis.com/en/pro-app/tool-reference/data-management/mosaic-to-new-raster.htm
[10] https://pro.arcgis.com/en/pro-app/tool-reference/image-analyst/how-raster-calculator-works.htm
[11] https://pro.arcgis.com/en/pro-app/help/analysis/geoprocessing/charts/histogram.htm
[12] https://pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/zonal-histogram.htm
[13] https://pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/region-group.htm
[14] https://pro.arcgis.com/en/pro-app/tool-reference/spatial-analyst/h-how-zonal-geometry-works.htm
[15] https://www.e-education.psu.edu/geog487/sites/www.e-education.psu.edu.geog487/files/image/lesson07/Pro/L7Data.zip
[16] https://www.wri.org/initiatives/forest-atlases#project-tabs
[17] https://data.globalforestwatch.org/documents/7af94f001fde4955a19907f7864aa9cf/about
[18] https://data.humdata.org/dataset/cod-ab-cmr?
[19] https://www.openstreetmap.org/
[20] https://www.globalforestwatch.org/map/?map=eyJjZW50ZXIiOnsibGF0Ijo4LjU5NDQ4MzAwNjE3NTk4LCJsbmciOjE1LjU4NzM5NjAxODk5MDEwNn0sInpvb20iOjQuNzQwMjk0NTg0NTg0ODQ4fQ%3D%3D&utm_campaign=treecoverloss2022&utm_medium=bitly&utm_source=GlobalForestReview
[21] https://colorbrewer2.org/
[22] http://www.globalforestwatch.org/
[23] https://landsat.gsfc.nasa.gov/
[24] https://earthexplorer.usgs.gov/
[25] http://pdf.wri.org/gfw_centralafrica_full.pdf
[26] https://pfbc-cbfp.org/forests-2008.html
[27] http://data.globalforestwatch.org/datasets/000078d77c404100be9d1ab027d1fa9e
[28] http://www.wri.org/blog/2014/02/9-maps-explain-worlds-forests
[29] http://www.wri.org/resources
[30] https://www.globalforestwatch.org/map/?map=eyJjZW50ZXIiOnsibGF0Ijo1LjkxODMyMjY1MDM1NzMwOSwibG5nIjoxMy4yOTcyNDUzMDIwMzIyNzV9LCJ6b29tIjo1LjI4ODM1MzUwMTUwMTcxLCJkYXRhc2V0cyI6W3siZGF0YXNldCI6ImludGVncmF0ZWQtZGVmb3Jlc3RhdGlvbi1hbGVydHMtOGJpdCIsIm9wYWNpdHkiOjEsInZpc2liaWxpdHkiOnRydWUsImxheWVycyI6WyJpbnRlZ3JhdGVkLWRlZm9yZXN0YXRpb24tYWxlcnRzLThiaXQiXX0seyJkYXRhc2V0IjoicG9saXRpY2FsLWJvdW5kYXJpZXMiLCJsYXllcnMiOlsiZGlzcHV0ZWQtcG9saXRpY2FsLWJvdW5kYXJpZXMiLCJwb2xpdGljYWwtYm91bmRhcmllcyJdLCJvcGFjaXR5IjoxLCJ2aXNpYmlsaXR5Ijp0cnVlfV19&mapMenu=eyJkYXRhc2V0Q2F0ZWdvcnkiOiJmb3Jlc3RDaGFuZ2UifQ%3D%3D
[31] http://www.youtube.com/watch?v=Oe1RYWBuhrE
[32] https://www.youtube.com/watch?v=pJD4_lwyy68
[33] https://www.youtube.com/watch?v=xHSRCeU1Hdc&feature=related
[34] https://youtu.be/a-aXuEU2Z4A