Undergraduate Remote Sensing Group Project for Dr Khan
VEGETATION REGROWTH RATES IN YELLOWSTONE AFTER THE 1988 FOREST FIRES
FALL SEMESTER OF 2017
University of Houston
Presentation Abstract
The Yellowstone National Park during drought seasons suffers from a high risk of wildfires which pose a threat to both local wildlife and vegetation. To explore potential solutions for stabilizing the park area, this project aims to look at the potential regrowth strength of two of the parks largest native tree populations, the Lodgepole and Whitepine. We examined regions in southeast Yellowstone near Yellowstone lake from the years 1987 before the fire to 2008 twenty years after the fire to best understand the damage to and recovery of these trees after the fire.
Objective of the Project
Fire prone regions of the United States are associated with regrowth risk such as permanent loss of vegetation, and the introduction of post fire invasive species which reduce the stability of local plant life. One solution to reducing regrowth risk within regions is the inclusion of native plants with a high post fire regrowth potential. In the case of this study which observes the Yellowstone National Park, and the regrowth of Lodgepole and White pine tree populations we hope to find a correlation between the two tree populations, and the rate of regrowth following the 1988 wildfire over a period of 20 years.
Program and Data
For the entirety of our project, we used ENVI 5.3 (64-bit) to process and examine our data.
​
Our Multispectral map images were taken by Landsat 5 and obtained from https://earthexplorer.usgs.gov/. The images ranged from the months of July-October from the years 1988 to 2008.
Methods
These are the processes we used to classify and compare data.
-
Initial calibrated data was obtained from the USGS EarthExplorer platform
-
Change detection differences run within Envi to utilize the band math function in order to subtract initial and final dates within the data
-
ROI’s established to allow local application of processes
-
Maximum likelihood classification run to grade vegetation based on burn severity
-
Change detection run to classify exclusively burned portions of the study area
-
ROI’s generated for burnt and unburnt areas specifically
-
Mask generated from burn area ROI’s
-
Maximum likelihood classification run to classify all portions of burned area within the study region
-
ROI to vector processing used to convert burned region ROI to vector data
-
Mask of burned area generated from vector data
-
NDVI run on both mask to determine changes within the data across the chronological study range
-
NDVI was chosen as the preferred classification method over NBR classification in preference of its ability to accurately discern vegetation as well as the methods specific application scope towards vegetation populations
Workflow Procedure
Overall processing was carried out across the ENVI and ENVI Classic platforms. Our data underwent various stages of calibration and which are depicted in the displayed image.
​
Picking Our Areas of Observation
Using maps taken from National Park Services' website, we researched the vegetation and fire history of Yellowstone National Park. From this site, we were able to gather information about the whitepine and lodgepole trees, as well as a map of the regions various fires affected. With these maps, we selected the whitepine and lodgepole areas that we would attempt to observe, indicated by the labeled red boxes on each map image. The coordinates for the areas is 44.4280 N, 110.5885 W.
Whitepine True Color Images
Processed true color images displaying the Whitepine population are displayed below, the images are ordered from oldest being the 1987 year to youngest represented by the 2008 year, moving in order of left to right.
(To best view these images, and the images following, it is recommended you click on them to enlarge them on your screen.)
Processed true color images displaying the Lodgepole population are displayed below, the images are ordered from oldest being the 1987 year to youngest represented by the 2008 year, moving in order of left to right
Lodgepole True Color Images
NDVI Classification Key
We used this classification scheme for both the whitepole and lodgepole NDVI images. Values between 0.2-0.4 could be grasslands, flowers, or growing saplings. Meanwhile, values 0.4-1.0 were considered healthy pine trees; being more dense and thriving the closer to 1.0 the pixel was.
Whitepine NDVI Images
Interpretation of collected Whitepine data
The collected Whitepine data after processing shows a growth trend that prefers areas within the park at higher elevations. The greening progression of tree growth in the above images depicts the growing path of the Whitepine tree population, one which can be seen to do well in the higher elevated regions of the park.
Lodgepole NDVI Images
The Lodgepole data shows a preference of growth outward from the major burn area. Due to the nature of the Lodgepole preferring to disperse its seeds during high temperature events such as wildfires there is a trend of growth originating outward from the burn area of the image. The increasing greening shown in the images displays the growing Lodgepole population over the years following the fire.
Interpretation of collected Lodgepole data
Conclusion
Final results and numerical data show a trend of overall growth across the twenty year study range in terms of both the Lodgepole and Whitepine tree populations. The sharp increase in lower NDVI values when extrapolated to the overall curve of the tree population behavior shows a net decrease in burned areas within the study region, and a net increase in healthy growing vegetation. Both tree populations as a result are concluded to be healthy sources of post fire regrowth stabilization, and have a natural tendency to grow well over large areas when in their native environments. Long term implementation of this research aims to conitnue monitoring the regrowth progress of both tree populations for a minimum of 20 more years to see if population numbers are capable of reaching their pre-fire quantities.
In Closing
Whitepine Results
Lodgepole Results
Using the Whitepine Pixel Count, we were able to create a graph representing the relevant data; namely the shrubs/grassland and dense forests. After the fires in 1988, the region is dominated by shrub and grassland, with dense whitepine forests recovering overall but never surpassing shrubs and grasslands again.
​
Similarly, the Lodgepole Pixel Count data was used to create a similar figure for the lodgepole trees. Just as with the whitepines, the lodgepoles are replaced by shrubs and grassland after the fires of 1988. Though the dense lodgepole forests increase, but also never reach the ratios of trees to grassland that they were in 1987 before the fires.
Results
Group Members
Group Members are:
Erin "Alaric" Doe
Christ Niamike
Wyatt Barrs
Justin Lindlof
Work Cited
Images were created using Envi 5.3 (64-bit)
Images were acquired by Landsat 5 at: https://earthexplorer.usgs.gov/
Maps of Yellowstone National Park Vegetation and Fire History were taken from https://www.nps.gov/yell/learn/nature/plants.htm and https://www.nps.gov/yell/learn/nature/firehistory.htm respectively