Deliverable 3.1: CVS will conduct a formal, quantitative review of the current EEP-mandated monitoring
protocol to determine whether some of the measurements are redundant relative to EEP needs so that
they might be dropped from monitoring requirements, thereby reducing monitoring costs.
SUMMARY:
The Carolina Vegetation Survey (CVS), in collaboration with North Carolina’s Ecosystem Enhancement
Program (EEP), has developed a vegetation monitoring protocol for restoration sites. A goal of this
protocol is to provide a consistent, efficient, and repeatable means for tracking performance of
individual planted woody stems. The protocol offers other benefits to the state, including compliance
with national standards (U.S. Federal Geographic Data Committee) for vegetation classification (National
Vegetation Classification Standard), scalability with other ecologically-driven protocols, and digitally
integrated into a sophisticated database.
CVS currently maintains data for woody stems from 662 permanent monitoring plots across 72
restoration projects. Dimensions measured annually for these stems include height, diameter at
decimeter height (DDH), and diameter at breast height (DBH). In addition, each stem’s coordinates
within a monitoring plot are recorded, along with the stem source (live stake, bare root, ball and burlap,
etc.). Annual updates are also recorded for stem vigor and damage observed.
In early 2010, EEP and several monitoring firms requested that CVS review the protocol to determine if
the number of stem dimensions measured could be reduced without loss of data utility. Because current
regulatory requirements specify minimum survival rates of planted woody stems, CVS examined survival
data for planted woody stems as functions of stem dimensions. The goal of this review was to develop a
statistical model to determine which predictor variables (stem height, vigor, DDH, DBH, etc.) were
essential and/or extraneous in terms of stem survival over project years. The modeling approach
chosen for this review was a logistic regression using the Generalized Linear Model (GLM), using stem
survival (1 = survived, 0=died) as the dependent variable. Specific results of this approach included:






Increasing DBH contributed to greater likelihood of survival for stems >5 cm DBH
Both larger DDH and greater height contributed to greater likelihood of survival for all stems
Combining height and DDH did not much improve model performance
Complex model (all possible predictor variables) shows enhanced performance over simple
model
Removing DDH from complex model results in little change in model performance
A model predicting annual survival rate for stems based on stem height alone performed nearly
as well as a model based on both stem height and DDH
These results suggest that stem height and DDH are comparable in terms of predicting stem survival
from year to year. Because measurement of DDH requires more time and effort than measurement of
stem height, the CVS has concluded that modifications should be made to the CVS-EEP protocol that
eliminates DDH measurements, simplifying woody stem monitoring.