CHANGES IN VEGETATION RELATED TO BEAR RANGES BY: AURORA HAGAN, JAIME NIELSEN, KRISTA TRENDA.

CHANGES IN V

EGETATI

ON

RELATE

D TO B

EAR

RANGES

BY : A

UR

OR

A H

AG

AN

, J AI M

E N

I EL

SE

N, K

RI S

TA

TR

EN

DA

OUTLINE

• Brief history on black bears

• Objectives

• Methods

• Results

• Discussion

• References

NATURAL HISTORY

Geographic Range: Alaska, Canada , NE United States, Rocky Mountains, Appellation Mountains, and Northern Minnesota

• Smaller populations found through out the US

Home Range in MN: The northern third of MN, coniferous and deciduous forest

Food Habits: Omnivores—grasses, nuts, berries, insects, fish, and human food

June and Lily

• Located in St. Louis County, MN

OBJECTIVES

• Classify vegetation in study area

• Visual vegetation change over time

• Relate analysis of vegetation to bear use

METHODS

ArcGIS 2010:

• Convert Lat/Long GPS points to UTM 1984 WGS 15N to match satellite imagery

• Separate shape files were created per bear, per month in order to relate them more directly to the time frames chosen for comparison

METHODS

ERDAS 2010/2011

• Layer stacked the 7 bands from the Landsat images to produce one pyramid layer per image

• Clipped the images to focus on the study area

• Unsupervised classification : 16-20 classes, 15 iterations, 0.95 convergence threshold

• Recoded to combine like classes

• Accuracy Assessment : Stratified Random--100 points, 10 points minimum per class

• Change detection • Oct 2008 to Sept 2011• July 2006 to July 2010• April 2010 to Aug 2010

RESULTS – 4 YEAR COVER TYPE COMPARISONJuly 2010

July 2006

RESULTS – GENERAL CHANGE DETECTION• Oct 2008 to Sept 2011

RESULTS – CHANGE IN COVER TYPE AND BEAR MOVEMENT• April 2010 to August 2010

DISCUSSION – PROBLEMS!

GPS points –

• 2008 – limited points over a 2 month period

• October 2010 – limited points for June

• Inconsistent collection intervals

Landsat imagery –

• Must have Level 1 downloadable data• Limited our ability to directly match GPS points to images

• Clouds are pretty, but not in my image

• Lack of 2009 images

DISCUSSION – PROBLEMS!

NAIP imagery –

• Files are HUGE• Need a large GB flash drive

• The images aren’t user friendly

ERDAS –

• Images were classified and recoded differently for July 2006 and July 2010• Caused an unusable change detection map

DISCUSSION – UNUSABLE CHANGE DETECTIONChanges in all classes from July 2006 to July 2010

DISCUSSION – WOULD HAVE DONE DIFFERENTPlanning

• Looked at the GPS data first to help decide time ranges

• Looked at what imagery from GloVis was Level 1

• Learn how to correct pixels of cloud cover

REFERENCES

• Special Thanks to Dr. Lynn Rogers and Sue Mansfield

• Thanks to Professor : Joe Knight, Teacher Assistants : Lian Rampi

and Jennifer Corcoran

QUESTIONS??

????

??