H18+V4+Southeastern+France

Lab 11 & 12 Supervised Classification

Lab 10 and Report 2 for Change Detection with SPEAR Tools

Landsat Lab 9 Spatial Convolution Filtering

Landsat Lab 9 Principal Component Analysis

Landsat Atmospheric Correction Southeastern France

MODIS Tile H18 V4 covers parts of Northern Spain, Northwestern Italy, Switzerland, Western Germany, and much of France. Vegetation change in Southeastern France is most likely due to numerous fires, some intentionally set, causing loss of life and destruction of property. The following links provide examples of news reports covering wildfires from 2003 through 2010, NASA satellite views of the area affected, and a pdf reviewing fires in Europe in 2006.

BBC - France Steps Up Fight Against Arsonists

Euronews - Summer Wildfires Hit France and Portugal

RFI - Fires Ravage South of France

Earth Observatory - NASA photos of fires in France

Forest Fires in Europe 2006

RFI - Army Starts Forest Fires near Marseille



The following three maps display Southeastern France for the Spring, Summer, and Fall seasons in 2010. The maps display the Normalized Difference Vegetation Index (NDVI) which helps to assess the health and vitality of vegetation in a region. Healthy vegetation indicates favorable environmental conditions; whereas, poor vegetation may indicate droughts, less productivity, or a lack of vegetation. Between the years of 2003 and 2010, there were numerous forest fires in Southeastern France. NDVI time series assist government agencies in assessing the damage as well as the recovery from the fires, while also preparing and monitoring vegetation productivity levels and control of future fire outbreaks. The map for Spring indicates healthy vegetation in the lower elevations, but also a lack of vegetation in the higher elevations of the Alps. The dark green areas indicate strong vigor and healthy production of vegetation. The map for Summer indicates widespread vegetation with numerous areas of healthy vegetation and very few areas of poor or a lack of vegetation. The map for Fall indicates mostly poor vegetation where the areas of foliage are beginning to die with the change of the season. There are numerous areas of unclassified data most likely due to cloud and/or snow cover.





1. Anthrome file: France03An has 121 lines, 188 columns, and 1 band. The MODIS Time Series file: TSFrance03An has 202 lines, 313 columns, and 144 bands.

2. There are more lines in the MODIS Time Series than the Anthrome file because the pixels are different sizes. There are 144 steps and 16 observations per year.

3. Shrinking the Anthrome pixels would cause a loss of data.



The 5 most dominant classes are: 1. 32 2. 26 3. 51 4. 25 5. 31

4. Percentage of area covered by the 5 most dominant classes: 60.51%.

5. Classes 25, 26, 31, and 32 follow the same trend in that as mean increases or decreases the coefficient of variation does the same; however, class 51 is the opposite.

6. It is possible to tell when the growing season begins because the mean increases and then decreases as the growing season comes to an end and occurs every year.

7. The phenology would be different if there were a lack of humans in the area in that there would be more forests and no villages or croplands.

8. Even though all classes indicate the growing season trend, Class 51 Populated Forests is more obvious.

9. I'm curious to know why all the classes increase for the growing season, decrease slightly, and increase again before the decrease for the end of the growing season. I'm wondering if this is due to a rainy season during the late summer or early fall months..

_

In 2003 and 2010, there was an outbreak of numerous wildfires across portions of Europe. In 2003, many countries experienced a devastating heat wave causing crop failure and dry vegetative conditions. Using MODIS Burned Area Products for 2003 and 2010, I compared the two years to see where and the extent of fires in France. Each year consisted of five months, May through September of burned area product images. These images were combined in ENVI using layer stacking. A mask was built and applied to only select the pixel data values between 0 and 365. The burn dates were then summed for all months. Band Math was used to find all values greater than zero, and then the count of each was then summed. After dividing the sum of the burn dates by the sum of the counts, the burned data pixels resulted. A histogram for each year is provided indicating burn dates. The year 2003 resulted in more fires during the summer months. The year 2010 appears to have had fires during early fall. Many of the news reports I have researched indicate fires were mostly in Southern France. I am currently looking for any reports that indicate fires were widespread across Northern France as displayed in the 2003 map. Further research will show whether or not these areas were the result of a fire or the possibility that due to the quick onset of extreme heat during 2003, there was a quick decline in the vegetation structure; however, it is quite possible these fires were intentionally set for agricultural purposes after the harvest of wheat or other grains.

_ These data are distributed by the Land Processes Distributed Active Archive Center (LP DAAC), located at the U.S. Geological Survey (USGS) Earth Resources Observation and Science (EROS) Center (lpdaac.usgs.gov).

Photo of forest fire in France retrieved from: @http://news.bbc.co.uk/2/hi/in_depth/photo_gallery/3106749.stm