Inventories Stategies for Acacia Koa Across Diverse Forest Types
Dr.Tom Baribault, Forest Solutions Inc., Hawaii, 2015-01-14

As one of the most economically valuable native Hawaiian tree species, Acacia koa (koa) has been heavily exploited across the islands. Remaining koa populations are constrained to a few regions, but diverse site conditions, land use history, and current vegetation cover disallow a single strategy for comprehensive koa forest inventory...



1. Summary

As one of the most economically valuable native Hawaiian tree species, Acacia koa (koa) has been heavily exploited across the islands. Remaining koa populations are constrained to a few regions, but diverse site conditions, land use history, and current vegetation cover disallow a single strategy for comprehensive koa forest inventory. This report will review the major forest types in which significant koa populations occur and present appropriate inventory options. The practicality of these methods has been demonstrated by successful inventory projects conducted in each forest type. The objective of this document is to equip land managers from government agencies or private entities with forest inventory methods known to be suitable for measuring most koa forests in Hawaii.



2. Forest classification

Structure and composition of Hawaiian koa forests vary along several axes, from closed to open canopy, monotypic to mixed species, and extensive historical land use to limited degradation. Effective inventory techniques exist for each forest class, but correctly identifying the type of forest is a critical first step in developing an efficient inventory protocol. We have encountered four broad classes of koa forest that necessitate different inventory approaches. A brief field visit is usually sufficient to define forest type; qualitative definitions have been satisfactory for most applications.

2.1 Open canopy, low or intermediate koa density

Koa forests with open canopy may be located in windward or leeward areas, but share land use history involving some combination of cattle ranching and timber extraction. Open canopy sites may support high koa densities, but low densities are more common. Koa canopies do not overlap on these sites; low juvenile density and heavy ungulate browsing cause a sprawling growth habit (Figure 1) typified by a short bole with numerous lateral branches. Readily available satellite or aerial imagery (e.g. Google Earth) is often adequate to distinguish and even to measure crowns of individual trees.

2.2. Mixed species closed canopy, low koa density

A second common koa forest type comprises multiple overstory species but is usually dominated by Metrosideros polymorpha (ʻōhiʻa).Stands may have low koa populations due to substrate age or high-grade koa logging. Although this condition may arise via different pathways, the resulting structure is a closed overstory in which most individual tree canopies intersect. From aerial imagery, it is not generally possible to differentiate koa from ʻōhiʻa, nor is it a simple matter to reliably identify individual canopies of either species, except for emergents.

2.3. Mixed species closed canopy, intermediate koa density

This third forest type is becoming increasingly rare, and may be found chiefly in protected areas that have not experienced historic grazing or logging activity. Species composition includes some mixture of koa and ʻōhiʻa, though usually the latter dominates. In the oldest mixed closed canopy forests with an appreciable koa component, it is often possible to visually identify large koa trees using aerial or satellite imagery. Total koa density shows some correlation to the number of individually visible canopies, but this relationship is usually weak.

2.4. Monotypic closed canopy

A final koa forest type consists of koa-dominated, often monotypic, stands that result from forest management activities, including scarification and tree planting (Figure 4). Although the total land area occupied by this forest type is relatively small compared to the others, this kind of active forest restoration is accelerating. Unlike natural forest, land managers often keep detailed records on initial stocking, growth, and population in restored koa stands. For planted stands, most of which were established within the last decade, aerial imagery can be used to track annual population change. Intensively planted koa occupies more than 1,000 acres on Hawaii Island; Forest Solutions and others establish new plantings on an annual basis.



3. Inventory methods

Inventory methods should be selected by forest cover type, koa density, average tree size, and overstory species composition (Table 1).

3.1. Plot 3P sampling

The most efficient inventory method for open canopy koa forests is a modification of 3P sampling in which the estimated unit is a plot rather than a single tree (P3P). In many areas, this approach is substantially improved by use of remotely sensed imagery. Aerial imagery (Figure 5) is used to estimate parameters of interest (e.g. koa saw timber volume, number of trees, total biomass, ecosystem composition) within a systematic sampling grid of defined polygons. Measurements of these plots in the field are then correlated to estimated values. Inventory with P3P can be conducted on vast areas for minimal cost; Forest Solutions has completed projects exceeding 30,000 acres.

3.2. Variable radius sampling

In closed canopy forests with higher stem density of target species (i.e. closed canopy koa-ʻōhiʻa or plantation koa), stand level parameters can be accurately derived by measuring select trees from point locations using an angle gauge (Figure 6).This method is the industry standard for forest sampling and timber sale cruising. The procedure can be optimized by adjusting sampling efforts to reflect the parameter of interest, for example by increasing the number of plots at which trees are simply counted, rather than measured, in order to improve the estimate of basal area.

3.3. Fixed area sampling

Although fixed area sampling is inherently less efficient, it is useful in closed canopy stands where koa is rare and can be improved by using handheld GPS and narrow (20’) strips. Where aerial imagery cannot identify target crowns, fixed area strip cruises rival the efficiency yet exceed the accuracy of variable radius sampling. When koa trees are very rare, detection can be a challenge both in the field and with remote sensing. A strip type fixed area inventory improves likelihood of detecting trees; advanced methods can incorporate elements of variable radius and 3P sampling to improve accuracy.



4. Planning and analysis


4.1. Inventory objectives

With each of the preceding inventory methods, it is important to clearly establish the cruise objective. All of the methods can be adjusted to emphasize parameters of interest, for example, number of habitat trees that meet a certain size criteria, total standing volume of saw timber, invasive species presence, or biomass fractions for carbon accounting projects.

4.2. Stratification

Structure and species composition in Hawaiian forests can be extremely diverse over short distances (Figure 7) due to the mosaic of geology, land use history, invasive species incursion, and stochastic colonization. scale spatial heterogeneity within a parcel drastically reduces accuracy and therefore utility of the inventory results for planning and management purposes. The best way to improve outcomes on parcels with multiple cover types is to aggregate areas—called strata—with similar vegetation, substrate, or land use. Strata should be defined with Geographic Information Systems (GIS) using a combination of field reconnaissance, survey layers, and remotely sensed data.

4.3. Accuracy

Standard statistical methods in forest inventory yield estimates of average parameters (e.g. number of trees, saw timber volume per area, total parcel biomass). The uncertainty associated with these average values is expressed as the standard error of the mean. Improving inventory accuracy requires an exponential increase in sampling effort, e.g. doubled sampling to reduce a standard error from 20% to 10%. We recommend a 10% to 15% standard error using a 90% confidence interval.

4.4. Allometry and demographics

Tree allometry changes radically among forest types and across the windward – leeward gradient. For inventories that require accurate standing volume assessment, generating a local volume equation (Figure 8) is a vital step in the protocol. Allometric equations govern the ultimate accuracy of each inventory method described here. An inventory may achieve a low standard error on basal area or tree population, but the final answer cannot be trusted without using an appropriate allometric equation. Each method is also capable of generating a diameter class distribution (Figure 9) from which parameters of interest measured at the individual level can be extrapolated to large areas.



5. Conclusions

Populations of Acacia koa on the Hawaiian Islands exist in several forest types and across diverse land holdings. Managing the species for habitat conservation, long term Hawaiian cultural usage, and high value timber production can be a realistic goal, but will require accurately assessing the current state of the resource. The inventory methods presented in this report serve as a foundation upon which to standardize koa inventory across state, and to coordinate activities among government and private forest management entities.



6. Forest inventory services

Do you have questions about koa inventory on your property? Forest Solutions has 18 years of experience in Hawaiian forestry, and we would be happy to discuss forest management options specific to your parcel.

Web: www.forestsolutionshawaii.com/
Phone: +1 (808) 776-9900



7. References

Asner, G.P., et al. 2011. Front. Ecol. Env.
Bastien-Henri, S. et al. 2010. For. Ecol. Manage. 260: 403-410.
Iles, K. 2013. A Sampler of Inventory Topics. Chapters 8, 11, 12, 13.
Mascaro J., et al. 2014. Bio. J. Linn. Soc.
News Archive
Forest Planning
and Project Management
Consulting and
Independent Audits
Land
Stewardship
Management
Information Systems