Week 6: Final project, An Unsupervised Classification of Lake Tahoe Basin Land Use/Land Cover

Hello all, this is my last post for a while since I won't be taking another GIS class until the fall next year. 
My task:

The Lake Tahoe basin has been subject to many fires over the past few years and the Forestry Department is interested in seeing how much of the south western portion of the basin has been affected in the past 20 years. However, the Forestry Department realized that they do not have LULC data for 1999. I have been tasked with performing a LULC analysis of the Lake Tahoe region in order to assist land managers with identifying the amount and pattern of forest within the California portion of the Lake Tahoe basin. I have been provided Landsat 7 data from 1999 in order to complete this analysis.

Imagery Utilized:

I used the provided landsat_img_nad27.img which contains an 8-bit Landsat 7 ETM image with 6 bands of the Lake Tahoe Basin and was taken 04-15-1999. Meaning that this image has very radiometric and spectral resolution. The image’s projected coordinate system is NAD 1927 UTM Zone 10N with a projection in Transverse Mercator. The image’s geographic coordinate system is GCS North American 1927 with a spheroid of Clarke 1866. The image was sourced through UWF’s GIS Drive however it could also be sourced through the USGS Global Visualization Viewer (GloVis).

Project Outcome:

I created one map that illustrated everything I needed to at once. With a little knowledge of mountainous landscapes, I created 8 categories for the image: bare land, coniferous forest, grass, mixed forest, scrub brush, space forest, snow and water. Everything is pretty much self-explanatory but I personally consider mixed forest a mix of trees, scrubs, grass and sometimes snow or sparse forest.

The area of my study area is 62487.78 Hectares. And the percentage of land cover is as follows:

•                    Bare Land: 8717.7 hectares, 13.9%

•                    Coniferous forest: 11444 hectares, 18.3%

•                    Grass: 1455.5 hectares, 2.3%

•                    Mixed Forest: 8091.3 hectares, 12.9%

•                    Scrub Brush: 6281.3 hectares, 10.1%

•                    Sparse Forest: 11025.7 hectares, 17.6%

•                    Snow: 7029.21 hectares, 11.4%

•                    Water: 8463 hectares, 13.5%

My concerns for this project are that the image of the area that is photographed is taken in April which is a time of year that still has snow on much of the mountains. So it is possible that the areas categorized may change in a few weeks to months. It is difficult to tell the difference between what could be dense scrub brush and grass, mixed forest and scrub brush, and barren land can sometimes be as white as snow. Not only that but the greenery is often sparkly forested or shrub brush may look like trees. Meaning that this is not entirely perfect as mountain landscapes are very complex, that I am not a professional and these are just the numbers I got doing my best work.

Another thing is that I chose this particular area because I started trying to classify the entire image and ran into several issues where urban was also snow, grass was also water, sage wasn’t really showing up as its own thing. So to avoid the urban area and the issues it presented I chose this chunk of land and made a partially true story. The area is under conservation and there have been many fires which has affected the land and caused fine sediments to drain into the lake due to small amounts of desertification caused by fires. The U.S. does seem to make a yearly GIS maps based on areas impacted by fire, conservation issues, protection plans, land use and many other subjects. I am sure that they have a 1999 LULC map of the area but I wanted this project to be all my work and not use their data as a crutch.

Map:

It is below and I hope that it is as self-explanatory as can be. The large mage on the left is my unsupervised classification of land use and land cover for Lake Tahoe’s basin. The inset map is to show the original true color landsat 7 image and the images as well as the study area’s location in terms of the United States Map because it is one thing to say Lake Tahoe but not everyone knows where it is located exactly or that it is in two states. The scale bar for my LULC image is to the bottom left because it didn’t look ok anywhere else and the north arrow is in the lake itself so it didn’t conflict too much with my map. I feel like I should have added the area of each classification but it just made the map look too clustered.