Monitoring, Continuous Quality Improvement and Adaptive Management

Monitoring is of vital importance to an adaptive management plan - knowing where you are and where you've been makes is easier to figure out where to go and how to get there. Continuous quality improvement goes hand-in-hand with monitoring and can be summarized as an intensive learn-by-doing approach. In a discussion prepared by Prof. Chad Oliver, both of these concepts are described below.

Pack Forest will develop a comprehensive monitoring and adaptive management strategy during 2002. It will address such issues as:

• Protocols for inventory data storage, entry and update;
• Statistical methodology for inventory data collection and update in conjunction with the Continuous Forest Inventory.
• Strategies to augment inventory with measurements that are required for scheduled treatments, i.e. regeneration surveys, pre- and post-PCT stand examination and commercial harvest cruises. These may be enough to supplement CFI measurements in order to maintain the inventory and calibrate the growth model. Inventory data will not be developed enough to evaluate growth rates at the stand level, but should be sufficient to monitor growth for groups of stands. Stands could be stratified as follows: those from planting to pre-PCT age; post-PCT to time of commercial thin treatment; and those from last thinning to final harvest). Further stratification will be possible as more data is collected and growth can be calibrated within the prior groups based on site, species, etc.
• Strategies for handling inventory data that is collected in different years. Forest inventory data is most useful when brought to a common "measurement date". Stand measurements from early in the planning cycle could be projected to the end of the cycle to create a common baseline for development of the next Landscape Plan.

______________________________________________________________________________

Monitoring

Monitoring entails:

1. Sampling of the results for the various target objectives. Various statistical procedures can ensure that the sampling is robust and efficient.
2. Comparing the sample results with the targeted (projected) results to determine how well the desired objectives were achieved.
   

Other fields have have developed detailed statistical analyses used in sampling and comparison that can be applied to forestry (e.g., Fiegenbaum, 1983).

The PackMonitoring.xls template spreadsheet can be used to record this monitoring (download the spreadsheet by clicking on PackMonitoring.xls, saving the zip file and then accessing the Excel file). There are six worksheets in this spreadsheet file:

A. Active Stands Projected: This worksheet is identical to the Active Stands worksheet of the Portfolio.xls spreadsheet file. It contains all anticipated operations and results for each stand to be treated during the first management cycle (2000 - 2005).

B. Active Observed: This worksheet layout is identical to the Active Stands Projected worksheet (described immediately above). It is used to insert observed information on each stand in the appropriate columns.

C. Active Difference: This worksheet layout is identical to the two worksheets immediately above; however, it compares the projected and observed value in each cell (above) and gives a zero if there is no difference, a value of the difference if the compared cells have numeric values, and a word (e.g., VALUE or REF) if the compared cells have letter values that are different.

D. All Stands Projected: This worksheet is identical to the All Stands worksheet of the Portfolio.xls spreadsheet file. It contains all anticipated conditions for all stands at the end of the first management cycle (2000 - 2005).

E. All Observed: This worksheet layout is identical to the All Stands Projected worksheet (described immediately above). It is used to insert observed information on each stand in the appropriate columns.

F. All Difference: This worksheet layout is identical to the two worksheets immediately above; however, it compares the projected and observed value in each cell (above) and gives a zero if there is no difference, a value of the difference if the compared cells have numeric values, and a word (e.g., VALUE or REF) if the compared cells have letter values that are different.

As an overall proposal, initial monitoring can be done in the steps described below. The monitoring should be done near the end of each management cycle in preparation for improving the analyses and updating the plan.

1. Visual stratification: A visualization of each stand generated through LMS can be compared with the respective stand. Each stand is ranked on how well the visualization resembles the actual stand. (This can be recorded in Column W of the Active Difference and All Difference worksheets of the Monitoring.xls spreadsheet.)

2. Sampling actively treated stands: Much of the needed sampling will be accomplished during operational activities. These activities include pre-sale timber cruising, post-operation inspection (usually following thinning), and regeneration surveys (following planting). The sampling will be used to fill in the Active Observed worksheet in the Monitoring.xls spreadsheet file.

3. Subsampling other stands: Data for treated stands can be transferred to the All Observed worksheet. All stands can then be stands by "visual fit" according to rank number in column W of All Difference worksheet. A sampling system can be determined to supplement the data from the Active Observed stands for each stratified group.

Continuous Quality Improvement (CQI)

CQI accepts the fundamental system property that there will always be variation within a population-and so between the projected and observed stands. CQI systematically reduces the variation by:

1. determining the largest variations (done in monitoring, above);
2. finding the largest causes of these variations;
3. correcting these causes (This step differs from the common and problem-causing practice of making compensations for these variations. Such compensations lead to greater variations and unmanageable systems.) This correcting the causes is partially incorporated into Adaptive Management, described below.
4. reprojecting a plan for the next management cycle, to be monitored, processed through CQI, and projected again.

Specifically, CQI will seek causes of variation by examining the sampled data-and comparing between the projected and observed data-in several ways.

1. For all stands and for the stands stratified by visual "fit", "pareto" diagrams and a "ranking" charts will be used to determine the largest variations.
2. The inventory and other data projected in LMS will be statistically compared with the newly collected data for each stand to be examined in detail.

3. The statistical analyses, diagrams, charts, and other "process control" techniques will be used to identify and correct the largest errors.

Adaptive Management

Adaptive management consists of two components:

1. Incorporating new information and correcting errors so the ability to project outcomes of management alternatives accurately is improved. This component is a continuation of the CQI described above.
2. Doing applied research on a variety of techniques to achieve to objectives, to ensure that the treatments and assumptions in the assumed relations between objectives and values are correct. Different stands (relatively small units) are treated in contrasting ways, with some stands being treated in ways that current science believes will not achieve the desired objectives. Consequently, some stands will fail to meet the targeted objectives; however, it is possible that current science will be proved wrong, and these treatments will prove to meet the objectives better.

This component of adaptive management can be done as "on-line" research, where parts of the forest are treated in these contrasting ways. It can also be done as "off line" research, where the contrasts are tested in experimental forests (field laboratories). In "off line" research, the most promising way is applied to the non-research forest.

For the first management cycle (2000 - 2005) of the Pack Forest Landscape Plan, the second component of Adaptive Management will not be applied as "on-line" research. The management cycle will be used to solidify the landscape plan procedures. It is possible that such "on-line" research will be planned and applied during future management cycles.

• Mason McKinley, Pack Forest - introduction
• Prof. Chadwick Oliver, College of Forest Resources - discussion.
• Literature Cited - Feigenbaum, A. V. 1983. Total Quality control. McGraw-Hill, Inc., New York.