14.7.3 Labor Cost
Overall, with the integration of the automated stratification method into NASS area
frame operations, new state area frames are being built with improved objectivity,
efficiency, and accuracy at a reduced cost. For example, the traditional Oklahoma
area frame was constructed in 4552 employee hours, while the hybrid Oklahoma
area frame required 1980 employee hours to complete. In general, the reduction in
labor costs for other states will be similar. States with more intensive agricultural
production may have more cost savings, due to less manual editing, while states with
less agriculture may cost relatively more due to a larger manual editing effort.
14.8 Conclusion
This chapter presents and assesses an automatic stratification method and its integration with a traditional stratification method for use operationally. The automatic
stratification method uses the CDL to derive percent cultivation and subsequently to
stratify US land cover. The results of the automatic stratification of NASS area frame
PSUs of Arizona, Georgia, Ohio, Oklahoma, and Virginia indicated that the automated stratification method was more accurate in determining US percent cultivation
in intensively cropped areas and weaker in low agricultural areas. The automatic
stratification method achieved higher accuracies in strata 11, 12, 13, and 20, while
the traditional method achieved higher accuracies in stratum 40. The differences in
the accuracies were statistically significant at a 95% confidence level.
To increase operational area frame accuracy further, this chapter introduces a
hybrid operational process that integrates the automated stratification method with
traditional area frame construction manual editing/review procedures. In this hybrid
operational area frame construction process, the CDL-based automated stratification
method is used as the primary stratification method. The existing state area frames
are revised by using automated stratification as the primary stratification tool instead
of starting over from scratch. The hybrid process results in significant stratification
accuracy improvements in high, medium, and low agricultural areas. From
2014–2017, new area frames for Alabama, Arizona, Georgia, Nebraska, New
Mexico, North Carolina, Oklahoma, South Dakota, Wisconsin, and Nebraska were
successfully constructed using the integrated operational process. It was found that
the hybrid process achieved 12–51% accuracy improvements, based on JAS
reported data, in the nine test states. A mean 10% accuracy improvement can be
directly attributed to the addition of the automated stratification method and an
additional mean 18% accuracy improvement can be attributed to the manual editing
and review procedures. The mean percent cultivated land of all strata for all states
became consistent with the stratum definitions after revision. In addition, stratum
homogeneity is significantly improved using the automated stratification method
with reduced standard deviations recorded for 22 out of 31 strata across the nine
14 Geospatial Land Use and Land Cover Data for Improving Agricultural Area. . .
295
Overall, with the integration of the automated stratification method into NASS area
frame operations, new state area frames are being built with improved objectivity,
efficiency, and accuracy at a reduced cost. For example, the traditional Oklahoma
area frame was constructed in 4552 employee hours, while the hybrid Oklahoma
area frame required 1980 employee hours to complete. In general, the reduction in
labor costs for other states will be similar. States with more intensive agricultural
production may have more cost savings, due to less manual editing, while states with
less agriculture may cost relatively more due to a larger manual editing effort.
14.8 Conclusion
This chapter presents and assesses an automatic stratification method and its integration with a traditional stratification method for use operationally. The automatic
stratification method uses the CDL to derive percent cultivation and subsequently to
stratify US land cover. The results of the automatic stratification of NASS area frame
PSUs of Arizona, Georgia, Ohio, Oklahoma, and Virginia indicated that the automated stratification method was more accurate in determining US percent cultivation
in intensively cropped areas and weaker in low agricultural areas. The automatic
stratification method achieved higher accuracies in strata 11, 12, 13, and 20, while
the traditional method achieved higher accuracies in stratum 40. The differences in
the accuracies were statistically significant at a 95% confidence level.
To increase operational area frame accuracy further, this chapter introduces a
hybrid operational process that integrates the automated stratification method with
traditional area frame construction manual editing/review procedures. In this hybrid
operational area frame construction process, the CDL-based automated stratification
method is used as the primary stratification method. The existing state area frames
are revised by using automated stratification as the primary stratification tool instead
of starting over from scratch. The hybrid process results in significant stratification
accuracy improvements in high, medium, and low agricultural areas. From
2014–2017, new area frames for Alabama, Arizona, Georgia, Nebraska, New
Mexico, North Carolina, Oklahoma, South Dakota, Wisconsin, and Nebraska were
successfully constructed using the integrated operational process. It was found that
the hybrid process achieved 12–51% accuracy improvements, based on JAS
reported data, in the nine test states. A mean 10% accuracy improvement can be
directly attributed to the addition of the automated stratification method and an
additional mean 18% accuracy improvement can be attributed to the manual editing
and review procedures. The mean percent cultivated land of all strata for all states
became consistent with the stratum definitions after revision. In addition, stratum
homogeneity is significantly improved using the automated stratification method
with reduced standard deviations recorded for 22 out of 31 strata across the nine
14 Geospatial Land Use and Land Cover Data for Improving Agricultural Area. . .
295