(10) Identify areas of CDL omission and commission error using the previously
mentioned ancillary data sources.
(11) Manually edit the original area frame PSUs using ESRI’s ArcGIS based on
the results of steps 8–10.
Automated Stratification Steps Continued (Blue Boxes in Fig. 14.9):
(12) Compute percent cultivation of each area sampling frame PSU in the newly
updated area frame as a final review step to identify PSUs in which the
automated stratification results do not match with the current stratum
definitions.
(13) Define strata based on the percent cultivation calculation. This step includes
conducting a final review of these nonmatching PSUs to determine the
appropriate stratum definition for each.
(14) Revised area sampling frame based on the hybrid method is complete.
14.6 Integration Results
This section presents and discusses the assessment results for traditional, automated,
and hybrid area frame stratification methods. The assessment metrics include:
(1) stratification accuracy; (2) mean stratum percent cultivation range, mean stratum
standard deviations, and mean stratum PSU size; and (3) area frame construction
labor cost.
14.6.1 Stratification Accuracy
The Alabama, Arizona, Georgia, Nebraska, New Mexico, North Carolina, Oklahoma, South Dakota Wisconsin, and Nebraska area frames were successfully revised
from 2014–2016 using the hybrid area frame construction method. Table 14.5
compares the frame stratification accuracies by state and method.
As shown in Table 14.5, an average 9% accuracy improvement is achieved based
on the automated stratification results with JAS reported data as in situ validation.
Oklahoma and Arizona have larger initial improvements using the automated stratification, while New Mexico, Georgia, and South Dakota have lower accuracy
improvements directly related to the automated stratification alone. On average, an
additional 18% increase in state-level accuracies is achieved with the addition of a
manual editing process to reduce PSU sizes and the manual review to identify areas
that are impacted by CDLs errors of omission and/or commission. An overall
average 28% area frame accuracy improvement for the nine new area frames
286
C. G. Boryan and Z. Yang
mentioned ancillary data sources.
(11) Manually edit the original area frame PSUs using ESRI’s ArcGIS based on
the results of steps 8–10.
Automated Stratification Steps Continued (Blue Boxes in Fig. 14.9):
(12) Compute percent cultivation of each area sampling frame PSU in the newly
updated area frame as a final review step to identify PSUs in which the
automated stratification results do not match with the current stratum
definitions.
(13) Define strata based on the percent cultivation calculation. This step includes
conducting a final review of these nonmatching PSUs to determine the
appropriate stratum definition for each.
(14) Revised area sampling frame based on the hybrid method is complete.
14.6 Integration Results
This section presents and discusses the assessment results for traditional, automated,
and hybrid area frame stratification methods. The assessment metrics include:
(1) stratification accuracy; (2) mean stratum percent cultivation range, mean stratum
standard deviations, and mean stratum PSU size; and (3) area frame construction
labor cost.
14.6.1 Stratification Accuracy
The Alabama, Arizona, Georgia, Nebraska, New Mexico, North Carolina, Oklahoma, South Dakota Wisconsin, and Nebraska area frames were successfully revised
from 2014–2016 using the hybrid area frame construction method. Table 14.5
compares the frame stratification accuracies by state and method.
As shown in Table 14.5, an average 9% accuracy improvement is achieved based
on the automated stratification results with JAS reported data as in situ validation.
Oklahoma and Arizona have larger initial improvements using the automated stratification, while New Mexico, Georgia, and South Dakota have lower accuracy
improvements directly related to the automated stratification alone. On average, an
additional 18% increase in state-level accuracies is achieved with the addition of a
manual editing process to reduce PSU sizes and the manual review to identify areas
that are impacted by CDLs errors of omission and/or commission. An overall
average 28% area frame accuracy improvement for the nine new area frames
286
C. G. Boryan and Z. Yang