Between Dec 19, 2024 and Jan 2, 2025, datasets can be submitted to DRUM but will not be processed until after the break. Staff will not be available to answer email during this period, and will not be able to provide DOIs until after Jan 2. If you are in need of a DOI during this period, consider Dryad or OpenICPSR. Submission responses to the UDC may also be delayed during this time.

Browsing by Subject "Forest health indicators"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Post-stratified estimation of Coarse Woody debris Volume using the down woody materials sample of forest inventory and analysis.
    (2010-05) Hatfield, Mark A.
    The Forest Inventory and Analysis (FIA) program of the USDA Forest Service conducts a nation wide survey of America’s forests. FIA field crews collect data on tree size, condition, and species, as well as data on the conditions in which they grow from a network of permanent ground plots known as Phase two plots (P2). FIA crews also collect more detailed forest health indicators, including data on Coarse Woody Materials (CWD), on a 1 16 subset of the P2 sample. This subset is known as the Phase 3 (P3) sample. FIA regularly publishes reports on the quantity and quality of America’s forests using data from the P2 sample. A post-stratified estimation technique is used increase the precision of the estimates without increasing the sample size. Currently, research on how to best apply the post-stratified estimator to produce estimates of the P3 forest health indicators has been lacking. This thesis will address this gap by testing 18 candidate geospatial layers (both categorical and continuous) as stratification layers to produce estimates of CWD volume in the Lake-states region of Minnesota, Wisconsin, and Michigan. Continuous geospatial layers will be broken into two to five strata using an optimization algorithm. A simulation experiment is used estimate the long term effectiveness of successful geospatial layers. The simulation experiment is performed to compare the conditional and unconditional variance estimators of the post-stratified estimators. Successful geospatial layers are then applied to sub-populations of varying sizes to determine the effect of spatial extent on the post-stratification method. Stratification layers derived from remote sensing products provided the best results. Using two or three strata is recommended because further partition of the population simply produces ineffective sliver strata. No difference was detected between the two competing variance estimators. The effect of spatial extent of the stratification was volatile. The use of large spatial extents is recommended. The conclusion of this thesis summarizes the lessons learned throughout as well as ideas for future research on the topic.