**Aspect 2: Designation of social-ecological zones (***wao/kai***) allowed for the management of population dynamics for key resource species across social-ecological regions (***moku***).**

Terrestrial social-ecological zones (*wao*) within a social-ecological region (*moku*) were designated by a two-word term beginning with "*wao*" and followed by an epithet that described their primary purpose and indicated appropriate activities within each zone [16] (Table 2, Figures 1 and 2). Social-ecological zones in the marine environment (*kai*) have been historically documented within this system [7,29] (Table 3), but these have yet to be comprehensively examined or explored with spatial modeling. Both *wao* and *kai* spanned across the *moku*, which effectively divided each individual social-ecological community (*ahupua*'*a*) vertically, while connecting it horizontally to adjacent *ahupua*'*a* within a *moku* (Figure 1). The vertical divisions allowed for system-based management within each *ahupua*'*a*, while the horizontal connections between *ahupua*'*a* allowed for coordinated management of the population dynamics of key resource species between *ahupua*'*a* within each zone spanning a *moku*. This was achieved, in part, by a rotating system of harvest restrictions (described below), which ultimately facilitated management for maximum cumulative abundance and benefit of the entire system—a point that is elaborated below (Aspect 3).

**Table 2.** The five terrestrial social-ecological zones (*wao*) that appear to have been recognized on the island of Kaua'i. Management implications for each zone are provided (based on Table 3 in Winter and Lucas [16]).



**Table 2.** *Cont.*

**Table 3.** An abridged list of select social-ecological zones (*kai*) within the marine environment as documented by Maly and Maly [29]. Translations of the meaning of these zones are provided by the authors.


**Figure 1.** A schematic model depicting the layout of a single social-ecological region (*moku*) including the structure of both social-ecological zones (*wao* and *kai*, designated horizontally) and of social-ecological community boundaries (*ahupua*'*a*, designated vertically) to convey the framework for the biocultural resource management of the *moku* system in the Hawaiian archipelago in the pre-contact period. This framework provided for management in both the horizontal and vertical dimensions. Social-ecological zones are based on those identified from the island of Kaua'i [16].

**Figure 2.** A spatial model depicting the layout for the social-ecological region (*moku*) of Halele'a on the island of Kaua'i, including the social-ecological zones (*wao* ) that dictated resource management in each social-ecological community (*ahupua*'*a*), as determined by Winter and Lucas [16]. Each *wao* is represented by a different color as indicated in the key. This *moku* contains nine *ahupua*'*a*, each of which are labeled here by name. Not all *ahupua*'*a* modeled here have all five *wao* documented from the island of Kaua'i, which indicates that each *ahuapua'a* had varying levels of access to and amounts of biocultural resources.

#### **Aspect 3: Population management of key biocultural resources operated on an ecoregion scale.**

*Moku* provide ideal units for examining management systems for key resources [31]. While they are often understood as political boundaries, their alignment facilitated decentralized resource management under *ali*'*i* '*ai moku*, the royal title for those who administered resources in a *moku*. *Moku* boundaries encompass land- and sea-scapes and are aligned with biophysical attributes of island ecosystems—such as landscape aspect, topography, climate regime, wave exposure, watershed classification, forest distribution, substrate type, and aquifer boundaries (Figures 3 and 4). In this regard, *moku* boundaries are more closely aligned with the scientific understanding of an archipelago-scale ecoregion than any other unit of land division recognized in pre-contact Hawai'i. Ecoregions are relatively large units of land containing a distinct assemblage of natural communities and species, with boundaries that approximate the original extent of natural communities prior to major land-use change [32]. While usually referred to on a global scale, we use this term on an archipelago scale. This concept is explored in more detail below.

Owing to Hawai'i's orthographically driven climate patterns across the landscape and shoreline, bio-physical resources—such as sunlight, rainfall, temperature and wave energy [33,34]—ultimately drive natural resource abundance and the potential for cultivating biocultural resources via agro-ecological and aquaculture systems. While there are climatic similarities across *moku*, there are also key differences between *moku.* These differences can be seen with an RGB visualization of equalized temperature (◦C), solar radiation (W/m2), and rainfall (mm) [35,36] respectively (Figure 3). This can also be visualized in data distributions in histograms of climatic and landscape variables island wide, across *moku*, and within social-ecological zones (Figure 4). The overlay of *moku* boundaries in Figures 3 and 4 reveal clear patterns of climatic similarity within each *moku*. This suggests these divisions optimized land uses and had the potential to contain specialized biocultural resources. In particular, *wao kanaka* zones (including coastal areas) are primarily differentiated between *moku* by solar radiation, rainfall, temperature, and wave energy. This suggests that human interaction with the environment in these areas helped to further distinguish the *moku* from one another and inform appropriate uses. This is evident in the varying forms and intensification of agriculture associated with each *moku* [8], as well as coastal resource development or extraction [29]. This research does not assume that only these physical variables strictly dictated *moku* or *wao* boundaries while disregarding social and cultural drivers; however, an examination of the patterns of both similarities and differences across these spaces does suggest a logical grouping of resource uses as dictated or limited by some bio-physical constraints. *Moku* boundaries also correspond well with the population dynamics of key biocultural resources—such as fish, birds, invertebrates, and plants—that could be more effectively managed in the context of their natural ranges, and in their respective gene pools within ecoregions. Specific examples of key species in these life-form categories are given below.

Fresh-, brackish-, and salt-water vertebrate and invertebrate species were important components of traditional food systems in pre-contact Hawai'i [29]. At the local (*ahupua*'*a*) and district (*moku*) levels, fishing activities and catch distribution were strictly disciplined by a system of rules and regulations—born out of an understanding about the life cycles of various aquatic species—that were embedded in socio-political structures and religious systems (discussed below). Harvest management was not based on a specific amount of fish, but on identifying the specific times and places that fishing could occur so as not to disrupt basic life-cycle processes and habitats of important food resources [37]. Many of these laws provided protection for important species and allowed Hawaiians to derive sustenance from the ocean for centuries [38]. Knowledge about fish habitat needs, behaviors, and life cycles paved the way for the development of various aquaculture technologies that both increased and stabilized the production of fish biomass [29,39] in the social-ecological system.

Watersheds that contained perennial streams flowing from the mountains to the sea were provided with important vertical dimensions of instream food resources in the form of various species of native fish ('*O*'*opu*) and macroinvertebrates ('*Opae ¯* and *H¯ıh¯ıwai*) (Table 4). '*O*'*opu* were the most commonly-referenced fish listed as a traditional food source by native Hawaiians on islands with perennial streams in the middle of the 19th century, which alludes to the importance of these freshwater protein sources in that era [29]. This was particularly true for families living inland from the coast. Hawai'i's native stream species are all amphidromous [40] in that they move out to sea as larvae and return to freshwater as sub-adults to complete their juvenile and adult phases [41,42]. For '*O*'*opu*, eggs are laid and fertilized in nests, often close to stream mouths. Newly hatched larvae passively drift with stream currents into nearshore areas as marine plankton [43], then metamorphose and recruit into streams as juveniles [44]. The recruiting '*O*'*opu* are known in Hawaiian as *hinana*, which is the first size class recognized as edible [39]. Adults of each species predictably distribute themselves into high densities along elevational zones in the stream continuum [45], where they may be reliably collected seasonally. Given their amphidromous life histories, sustaining native '*O*'*opu*, '*Opae* (an ethnogenus comprising *Atyoida* and *Macrobrachium*), and *H¯ıh¯ıwai* (*Neritina granosa*) larval production from streams within and among watersheds is important to replenish oceanic planktonic populations as cohorts mature to enter streams as juveniles. An ecoregional-scale of resource management, consisting of multiple adjacent streams combined into an ecoregion management unit (*moku*) would, therefore, serve to optimize larval production regionally and be beneficial in sustaining native food resources in streams on all islands.

**Figure 3.** A visual interpretation of climate as delineated by histogram-normalized color combinations of red, green, and blue to simultaneously visualize gradients and combinations of temperature, solar radiation, and rainfall (red: mean annual temperature (◦C); green: mean annual solar radiation (W/m2); blue: mean annual rainfall (mm)). Social-ecological region (*moku*) boundaries (thick black lines), and social-ecological zone (*wao*) boundaries (thin dashed lines) representing the data produced by Winter and Lucas [16] are overlaid atop the island of Kaua'i. All climate data are from Giambelluca [35,36]. Areas with blue dominance represent relative rainfall abundance, areas of green dominance represent relative solar radiation abundance, and areas of red dominance represent relative warmer temperatures. This results in color mixes that demonstrate these climatic variables, with the Venn diagram providing a color key for visual interpretation of the mean annual climatic variability.

**Figure 4.** Histograms of climate and landscape variables (columns) for the example island of Kaua'i. From left to right: mean annual rainfall (mm), mean annual temperature (◦C), mean annual solar radiation (W/m2), long-term wave power (Kw/m), and landscape aspect. Rows display island-wide data distribution (bottom) and subsets of socio-ecological zones. Grey histograms represent all data in the zone or island with color coordinated distribution lines display distribution of each according to *moku*. Base-layer image of Kaua'i indicating social-ecological zones is from Winter and Lucas [16]. The boundaries of each of the five *moku* (Halele'a, Ko'olau, Napali, Kona, and Puna) for Kaua'i are ¯ indicated in separate colors.

Nearshore fish species were also important as a protein source, particularly for people living along the coast, and were managed on an archipelago-based ecoregion scale for abundance [29]. Management tools included the use of temporal and seasonal closures, a practice widely used in traditional Pacific marine tenure systems. Such closures most often applied to reduce intensive harvest of spawning fish or aggregations that occurred during lunar, seasonal, or annual cycles [4,46]. A number of pelagic and migratory species were heavily relied on as food sources, and effective management of their populations was more appropriately addressed at the *moku* level. An example of such management is evident in the ancient fishing regulation of '*Opelu ¯* (mackerel scad, *Decapterus* spp.)—in the *moku* of Kona Hema, Hawai'i Island, which happened beyond the seaward boundary of the *ahupua*'*a* in that ecoregion. This regulation mandated that '*Opelu ¯* be actively fed (*hanai ¯* '*ia*) in their natural aggregation areas (*ko*'*a*) during the restricted (*kapu*) season, which was associated with their spawning period. Each fishing family had a designated *ko*'*a* to *hanai ¯* during the *kapu* season. If they fulfilled that responsibility they were allowed to fish within any of the *ko*'*a* during the unrestricted (*noa*) season, after first harvesting from the one they tended. If, however, a family did not fulfill their responsibility to *hanai ¯* their designated *ko*'*a* in the *kapu* season, they then lost their privilege of fishing for '*Opelu ¯* in the following *noa* season. This is recalled in the proverb, "*Hanai a ¯* '*ai*", [29] that roughly translates to, "Feed [the fish], and [you may] eat", (translation by authors). Regulations that restricted the fishing of key species during their spawning season and calling for the active feeding of them during this period likely increased the fecundity of key resource

fish species for the entire *moku*. The six-month *kapu* season for '*Opelu ¯* was the *noa* season for *Aku* (skipjack tuna, *Katsuwonus pelamis*), a predator of juvenile '*Opelu ¯* [29], therefore this restriction/feeding season for '*Opelu ¯* corresponded with a shifted dietary reliance of Hawaiians to top-predator species as a protein source. As such, in addition to limiting pressure on key lower trophic level fish species, harvesting their predators reduced their natural mortality. When the *kapu* was lifted for '*Opelu ¯* fishing, the six-month *kapu* for *Aku* fishing commenced [7,29,39], thus allowing for population recovery of that species. The rotating *kapu*/*noa*, *noa*/*kapu* seasons alternated between these two species on an annual basis. Another important nearshore fish, '*Anae holo* (striped mullet, *Mugil cephalus*), was a prized species that migrates along coastal areas and into estuaries within an archipelago-scale ecoregion, and was a focal species in aquaculture systems. Not only were '*Anae holo* fished for as they passed through the coastline of an *ahupua*'*a*, they were also attracted into aquaculture systems, which were designed to create or enhance habitat for key resource species in a contained area. This included six classes of fishponds [29,39]. The replenishment of fishponds was dependent on the spawning success of this and other species, which happens on a scale that is more closely aligned with *moku* boundaries than any other scale of land division in ancient Hawai'i.

Birds—including forest birds, waterfowl, seabirds, and other migratory species—were another key biocultural resource group as a source of both food for sustenance, and feathers for adornment. As with pelagic and migratory fish, the population dynamics of native birds extended beyond *ahupua*'*a* boundaries. Hawaiian honeycreepers (Fringillidae: Drepanidinae), a highly diverse passerine group relied upon for their feathers, can have home ranges of up to 12 ha [47]. In the context of inland forest at or near the apex of *ahupua*'*a* home ranges of native honeycreepers could most certainly go beyond *ahupua*'*a* boundaries, while staying well within the social-ecological zones (Figure 2) that spanned multiple *ahupua*'*a*—such as the *wao akua* and the *wao nahele ¯* in the case of forest birds. The *Koloa* (Hawaiian duck, *Anas wyvilliana*), once an important source of food associated with the *wao kanaka ¯* zone [8], has been documented to fly between wetland systems in the same *moku* [48]. Ground-nesting seabird colonies—such as those of the *'Ua'u* (Hawaiian petrel, *Pterodroma sandwichensis*), which was another food source when abundant—encompass the upland forest of entire *moku.* An example of this is the colony at Honoonapali [ ¯ 49]—the region of montane cloud forest encompassing the entire *wao akua* zone in the *moku* of Napali on the island of Kaua'i. Therefore, given that key resource birds have ¯ home ranges and population dynamics, which existed in social-ecological zones that spanned across many *ahupua*'*a* yet remained within *moku* boundaries, managing their populations for abundance would have been more effective if done at the *moku* scale.

Species ranges and population dynamics of native plants—as opposed to cultivated crops—were also not limited to *ahupua*'*a* boundaries. Native plants co-evolved with three natural vectors of dispersal—wind, birds (either internally or externally), and ocean currents. Coastal plants tend to be distributed by ocean currents, whereas inland species tend to be distributed by wind or wing [50]. '*Ohi ¯* '*a lehua* (*Metrosideros polymorpha*), the native tree with the highest biocultural value [51], has wind-born seeds that can be dispersed great distances. As for culturally-important trees with fleshy fruits—such as *Mamaki ¯* (*Pipturus* spp.), '*Alahe*'*e* (*Psydrax odorata*), and many others—avian dispersers are critically important, and such birds are responsible for the structure and diversity of forests in Hawai'i [52]. Therefore, diversity of culturally-important native plants, as well as the structure of forests depended on physical and ecological factors that existed on a scale more closely aligned with those of the *moku* than any other scale of land division in ancient Hawai'i.

The abundance of biocultural resources, needed by stewards of the *ahupua*'*a* for their sustenance and well-being, depended on ecological factors, including life cycles of key resource species, that operated on scales larger than that associated with *ahupua*'*a* boundaries. This makes the larger *moku* a more practical unit for management.

#### **Aspect 4: Ensuring high levels of biodiversity resulted in resilient food systems.**

Hawaiians in the pre-contact era used taxonomy to attribute names to specific units of biodiversity in their social-ecological system [25], which provided a means to manage the components at the foundation of a diverse range of sociocultural traditions. The management of biocultural diversity has been identified as an important aspect of maintaining—and potentially restoring—the structure, function, and resilience of social-ecological systems [21]. The same concept can be applied to food systems. There is a word in the Hawaiian language for famine—*w¯ı* [9]—which indicates that food was not perpetually abundant in all areas. Periods of famine are noted to have followed natural disasters, such as hurricanes, or climatic shifts which resulted in extended periods of drought [53]. This evidence suggests occasional short-term declines in food abundance, yet points to the importance of biodiversity for resilience of the food system. Some species of plants are referred to as "famine foods" [8,54], and the same is true for some species of marine life [39]. Resource managers had to maintain high levels of biodiversity (Table 4) throughout the social-ecological system as a means to facilitate resilience in the food system. Resource managers had tools to maintain abundance and biodiversity in the food system. These tools included various types of *kapu*, or harvest/access restrictions, to allow for the recovery of populations of key species [29]. When certain species had *kapu* placed upon them, many others in the system could be relied upon as substitutes—as indicated in the alternating *kapu* between '*Opelu ¯* and *Aku* (discussed above). The high levels of redundancy in wild food sources is indicative of a resilient food system, one that identified food sources that were relied on primarily in periods of scarcity.



#### **Aspect 5: Rotations of harvest restrictions were tools to manage for abundance of biocultural resources.**

Maly and Maly [29] comprehensively documented Hawaiian fishing traditions from the pre-contact era, through the Kingdom period, and into the territorial period—based on a compilation of historical records and oral histories. They documented rotating harvest restrictions (*kapu*) that were placed and lifted (making an area *noa* or free from restriction) on either a regular or intermittent basis. The Hawaiian biocultural resource management system employed various kinds of harvest and access restrictions (*kapu*). The punishment for breaking a *kapu* was swift and severe [7,22]. A summary of the types of *kapu* employed in Hawaiian biocultural resource management strategies is described below (Table 5). These various kinds of *kapu* were employed in concert with each other—on both a temporal and spatial scale—to manage for the long-term abundance of key biocultural resources, while at the same time ensuring that local communities could access resources for their daily survival and well-being. The process for deciding which kind of *kapu* to employ and when, with the goal of managing population dynamics within a *moku*, was done by implementing a multi-criteria decision-making process—such as that which is described below (Aspect 6).


