Graduate Thesis Or Dissertation

 

Quantifying Forest Mortality with the Remote Sensing of Snow Public Deposited

https://scholar.colorado.edu/concern/graduate_thesis_or_dissertations/cz30pt03v
Abstract
  • Greenhouse gas emissions have altered global climate significantly, increasing the frequency of drought, fire, and pest-related mortality in forests across the western United States, with increasing area affected each year. Associated changes in forests are of great concern for the public, land managers, and the broader scientific community. These increased stresses have resulted in a widespread, spatially heterogeneous decline of forest canopies, which in turn exerts strong controls on the accumulation and melt of the snowpack, and changes forest-atmosphere exchanges of carbon, water, and energy. Most satellite-based retrievals of summer-season forest data are insufficient to quantify canopy, as opposed to the combination of canopy and undergrowth, since the signals of the two types of vegetation greenness have proven persistently difficult to distinguish. To overcome this issue, this research develops a method to quantify forest canopy cover using winter-season fractional snow covered area (FSCA) data from NASA's Moderate Resolution Imaging Spectroradiometer (MODIS) snow covered area and grain size (MODSCAG) algorithm. In areas where the ground surface and undergrowth are completely snow-covered, a pixel comprises only forest canopy and snow. Following a snowfall event, FSCA initially rises, as snow is intercepted in the canopy, and then falls, as snow unloads. A select set of local minima in a winter FSCA timeseries form a threshold where canopy is snow-free, but forest understory is snow-covered. This serves as a spatially-explicit measurement of forest canopy, and viewable gap fraction (VGF) on a yearly basis. Using this method, we determine that MODIS-observed VGF is significantly correlated with an independent product of yearly crown mortality derived from spectral analysis of Landsat imagery at 25 high-mortality sites in northern Colorado. (r =0.96 ±0.03,p =0.03 ). Additionally, we determine the lag timing between green-stage tree mortality and needlefall, showing that needlefall occurred an average of 2.6 ± 1.2 years after green-stage mortality. We relate observed increases in the VGF with crown mortality, showing that a 1% increase in mortality area produces a 0.33 ± 0.1 % increase in the VGF.

Creator
Date Issued
  • 2015
Academic Affiliation
Advisor
Committee Member
Degree Grantor
Commencement Year
Subject
Last Modified
  • 2020-01-27
Resource Type
Rights Statement
Language

Relationships

Items