Typically, the forest planning hierarchy has three levels: 1) operational, 2) tactical, and 3) strategic. Temporal and spatial scales along with policy and objectives are used to define each level. Operational plans usually cover 1-5 years for watersheds and landscape units, and have detailed objectives and constraints. Tactical plans are applicable for 20-40 year planning horizons, and are generally applied at the landscape unit scale with less operational detail. Strategic plans cover long time horizons and are applied at the forest estate level to explore policy and set long-term strategies. These are generic definitions that can easily change, depending on ownership objectives, policy, and the nature of the forest. For example, it may be necessary to make long-term forecasts of the tactical plan in order to have confidence that the first 20-40 years are sustainable. Regardless of how each level is defined, the planning hierarchy must provide adequate linkages between all levels so that the operational plan is consistent with the tactical plan, which in turn is consistent with the strategic plan. Clearly, the linkages have to go both ways so the reverse is true.

Forest management has undergone radical change in the last decade that has created a plethora of planning initiatives to cope with new and conflicting objectives. Rarely, if ever, are these initiatives coordinated within a well designed planning hierarchy, so we tend to confound objectives and constraints, use inadequate or excessive spatial data, use inappropriate geographic scales, and have poor linkages between strategic and operational plans.

A properly designed hierarchy helps to establish the spatial resolution that is needed to meet objectives at each level. At the operational level, we must be confident that the harvest units are correctly engineered, meet short-term market demands, satisfy current policy and be reasonable accessible. These harvest units are always manually designed because the need for accuracy and detail cannot be confidently modeled. At the tactical level, harvest units can be manually planned from maps and aerial photographs, or they can be generated with computer models that attempt to capture fundamental design principles such as operability, opening size, timber types. While the manual design offers the greatest confidence in modeling current practices and local preferences, it is expensive and time consuming when applied to large areas. It is also a rare event when these manually planned harvest units match the engineered units developed for actual harvest. Differences can be traced to changes in objectives, policy, logging systems, and additional field constraints that evolve over time. However, there still remains a greater comfort level when using manually planned units in the tactical plan to ensure that what is proposed in the operational plan remains feasible in the future.

The argument for manually planned units has also been extended and applied to strategic plans. The ATLAS/SIMFOR Project has used the manual design approach for harvest units at both the tactical and strategic levels, and entire Timber Supply Areas have been manually blocked at great expense (using maps, photos, and ground checks). Preparations of spatial databases using manually designed harvest units is the largest expense and most serious bottleneck in our regional planning initiative.

There are many approaches to the blocking problem within spatial forest planning. Glen Jordan at UNB uses simulation to first create a volume-based strategic plan, then attempts to allocate the harvest to spatial blocks. Mark Jamnick, while at UNB used a similar approach, but preferred optimization to simulation. ATLAS has used a blocking-first approach, followed by harvest scheduling with simulation. TELSA (Klenner and ESSA) also uses a blocking-first approach (including tessellation), followed by stochastic simulation to examine a range of outcomes associated with natural disturbances. OPTIONS (DR Systems) and COMPLAN (Sims Reid Collins) both use a blocking-first, simulation-second strategy. We now see FSSim (BCMOF) moving in this direction as the Timber Supply Branch begins introducing spatial resolution to their model. Other approaches include simultaneous blocking and scheduling (FRENZY - Lockwood, and FSOS - Hugh Hamilton). This short survey shows that there are many ways to solve harvest scheduling problems, and a more detailed review will show that each has its advantages. Models and solution strategies reflect the thinking of their designers. If analysts/planners tend to think like the designer, they are attracted to that model and the solution strategy it employs.