thumbnail

Interpretation of high-resolution imagery for detecting vegetation cover composition change after fuels reduction treatments in woodlands

Ecological Indicators

By:
, , , , and
DOI: 10.1016/j.ecolind.2014.05.017

Links

Abstract

The use of very high resolution (VHR; ground sampling distances < ∼5 cm) aerial imagery to estimate site vegetation cover and to detect changes from management has been well documented. However, as the purpose of monitoring is to document change over time, the ability to detect changes from imagery at the same or better level of accuracy and precision as those measured in situ must be assessed for image-based techniques to become reliable tools for ecosystem monitoring. Our objective with this study was to quantify the relationship between field-measured and image-interpreted changes in vegetation and ground cover measured one year apart in a Piñon and Juniper (P–J) woodland in southern Utah, USA. The study area was subject to a variety of fuel removal treatments between 2009 and 2010. We measured changes in plant community composition and ground cover along transects in a control area and three different treatments prior to and following P–J removal. We compared these measurements to vegetation composition and change based on photo-interpretation of ∼4 cm ground sampling distance imagery along similar transects. Estimates of cover were similar between field-based and image-interpreted methods in 2009 and 2010 for woody vegetation, no vegetation, herbaceous vegetation, and litter (including woody litter). Image-interpretation slightly overestimated cover for woody vegetation and no-vegetation classes (average difference between methods of 1.34% and 5.85%) and tended to underestimate cover for herbaceous vegetation and litter (average difference of −5.18% and 0.27%), but the differences were significant only for litter cover in 2009. Level of agreement between the field-measurements and image-interpretation was good for woody vegetation and no-vegetation classes (r between 0.47 and 0.89), but generally poorer for herbaceous vegetation and litter (r between 0.18 and 0.81) likely due to differences in image quality by year and the difficulty in discriminating fine vegetation and litter in imagery. Our results show that image interpretation to detect vegetation changes has utility for monitoring fuels reduction treatments in terms of woody vegetation and no-vegetation classes. The benefits of this technique are that it provides objective and repeatable measurements of site conditions that could be implemented relatively inexpensively and easily without the need for highly specialized software or technical expertise. Perhaps the biggest limitations of image interpretation to monitoring fuels treatments are challenges in estimating litter and herbaceous vegetation cover and the sensitivity of herbaceous cover estimates to image quality and shadowing.

Geospatial Extents

Additional Publication Details

Publication type:
Article
Publication Subtype:
Journal Article
Title:
Interpretation of high-resolution imagery for detecting vegetation cover composition change after fuels reduction treatments in woodlands
Series title:
Ecological Indicators
DOI:
10.1016/j.ecolind.2014.05.017
Volume
45
Year Published:
2014
Language:
English
Publisher:
Elsevier
Contributing office(s):
Southwest Biological Science Center
Description:
9 p.
Larger Work Type:
Article
Larger Work Subtype:
Journal Article
Larger Work Title:
Ecological Indicators
First page:
570
Last page:
578
Country:
United States
State:
Utah