Baseline and monitoring studies of Alakahi, and Onomea streams, Hamakua Coast, Hawaii.

The Hawaii Stream Research Center

University of Hawaii (Center for Conservation Research and Training)

Michael H. Kido

Final Report

for

The Hawaii Tropical Botanical Garden

January 31, 1998

Project Background and Logistics, Physicochemical Comparisons, and Benthic Studies

Michael H. Kido

Background

This study was conducted by personnel associated with the Hawaii Stream Research Center (HSRC)(University of Hawaii) and originated in response to concerns raised by the US Fish and Wildlife Service (USFWS) and the State Division of Aquatic Resources (DAR) about the impacts of two unpermitted diversions on Onomea and Alakahi Streams, installed by the Hawaii Tropical Botanical Gardens (HTBG) in 1979 and 1989 respectively. Removal of these two diversions was required by the Hawaii Board of Land and Natural Resources in early September 1997. Additional issues included the impact on streams of HTBG's planned expansion in the area and their proposed future water withdrawals from Onomea and Alakahi streams.

Our purpose in this study was to assist the HTBG, as well as State and Federal agencies, in developing responsible long-term stream and watershed management plans for lower Onomea/Alakahi (Fig. 1-2) streams by: 1. collecting baseline data on stream biological quality (as evaluated by metrics including fish, invertebrate, and algal abundances and composition); and 2. establishing a stream monitoring program to provide a basis for evaluating variation in functional processes within- and between-streams over time.

It was not the purpose of the HSRC to intercede in legal or management decisions or recommendations made by State or Federal agencies regarding HTBG activities. Our intent was rather to work within management guideline requirements imposed by these agencies to provide needed scientific data and to assist the HTBG during the study period to comply with required stream protection and environmental mitigation efforts.

Study Area

Four streams (Kalaoa, Hanawi, Alakahi, and Onomea) discharge into Onomea Bay between Lae O Puni and Hokeo Point in close proximity to each other (Fig. 3).

Figure 3. Study sites on Onomea and Alakahi Streams on the slopes of Mauna Kea, Hawaii Island.

Alakahi and Onomea streams flow through the HTBG property and are thus directly influenced by Garden activities. These two streams are the smallest in the complex based on estimated stream length (4.2 km and 2.3 km respectively). Onomea originates at around 299 m and Alakahi around 229 m elevation (Fig 3). These two streams, therefore, have relatively small drainage areas. Hanawi, in comparison, has its origins near 914 m elevation and travels 13.2 km to the ocean. Kalaoa stream originates at about 348 m elevation and travels a distance of 4.7 km to Onomea Bay. The four streams enter Onomea Bay within a distance of each other of about 0.6 km of coastline and are within 1.6 km of each other at an elevation of around 244 m elevation (Fig. 3).

Previous Studies

In the Hawaii Stream Assessment (HSA)(1990), only Alakahi and Hanawi streams are rated for "Aquatic Resources Status". Alakahi Stream was reported as having "substantial" aquatic resources; however, no survey date or other biological resource data is listed. Hanawi was rated as "Outstanding" with the presence of all native indicator stream taxa. Timbol and Maciolek (1978) classify both Alakahi and Onomea as "Limited-consumptive", continuous streams that were diverted and extensively crossed by roadways but not channelized. The Hilo staff of the State Aquatic Resources Division (DAR) surveyed Alakahi and Onomea streams from their river mouths to the Old Mamalahoa Highway using the "DAR point-quadrat method" in March 1996. The DAR survey found that while native stream fishes were present in both streams, their populations were not robust. The most common native fish species in Alakahi stream was Awaous guamensis ('o'opu-nakea) while Sicyopterus stimpsoni ('o'opu-nopili) was reported as being more abundant in Onomea stream. No native invertebrates were reported in the survey; however, the introduced Tahitian prawn, Macrobrachium lar, was reportedly very abundant in both streams.

Study Rationale

The proximate nature of Alakahi and Onomea streams necessitates a "cross-ecosystem analysis" (Kelly and Levin, 1986) to define local ecology. Although few "cross-stream" comparisons have been undertaken in Hawaii, data on "between-stream" relationships provide useful comparisons of stream functional processes, an assessment of regional stream quality, and information needed for conservation planning and managment. For Alakahi and Onomea streams, therefore, this study was designed to provide baseline data on resident aquatic species and compare key physical and biological attributes/processes in these streams as a means of assessing biotic variation (natural vs man-induced) in these proximate streams over time.

Study Objectives:

For the study period in Onomea and Alakahi Streams:

1. Assess and compare within- and between-stream variance in physicochemical attributes (slope, sinuosity, stream discharge, substrate composition and particle size, temperature, pH, dissolved oxygen, and conductivity).
2. Create a baseline list of resident aquatic species.
3. Assess and compare processes affecting the abundances of algal and invertebrate communities.
4. Assess and compare relative abundances/composition of native and alien stream macrofauna.

Materials and Methods

Physicochemical Comparisons

Because the presence of the concrete diversions in both Onomea and Alakahi Streams was the primary impetus for initiation of this study, sites were established on both streams to include these structures. Both streams had steep waterfalls below the diversions and these drop-offs were used to delineate the beginnings of the 100 m length study sites. Sites were situated at nearly identical elevations (39 m elevation for Alakahi and 38 m elevation for Onomea) and both streams passed under Mamalahoa Highway (Alakahi through a culvert and Onomea under a bridge). The similarities in physical setting presented an ideal situation for comparing physical and biotic attributes/processes within and between these proximate Hamakua coast streams.

Palmer et al. (1997) present a convincing case for the importance of variance in habitat structure as an appropriate ecological metric when assessing biotic processes in streams. In this study, we follow Palmer et al.’s (1997) use of coefficient of variation to quantify spatial variability. To collect spatial data for an assessment of within-stream variance in physical habitat characteristics, both 100 m study sites were measured and flagged at 5 m intervals. Physical variables were measured at these intervals and summed among intervals to provide a mean and standard deviation for that variable within-streams. Between-stream comparisons were then made possible using these data. Incremental rise in slope for each 5 m linear upstream-distance was measured with a 10X hand-level and measuring rod (1.6 m length). Sinuosity, or the angular change in channel flow, was measured using a compass by estimating the direction of flow (degree change from North) at each 5 m increment, sited on a stationary measuring rod (1.6 m length) positioned at the stream’s edge. Mean flow (cms) and depth (m) at each 5 m increment were measured in 0.25 m intervals across the stream using a Swoffer flow meter and top-setting wading rod. Dissolved oxygen (mg L), % oxygen saturation, conductivity (mS/cm), pH, and temperature (deg C) were measured in situ (center stream and equidistant from both banks) using a HYDROLAB Datasonde 3 multiprobe with Scout2 data display.

Mean and variance in substrate particle size and % exposed surface was estimated at each 5 m increment (M. Kido unpublished method) using a sq m grid delineated across the stream channel by a secured meter-flagged line (see Kido 1997a). This grid was also used to sample the benthic environment (described below). In each sq m cell, a PVC frame was used to delineate a 1/6 sq m "major subunit" and the percent of out-of-water substrate (% exposed boulder) in each "major sub-unit" was estimated. "Major sub-unit"estimates were subsequently combined for each sq m cell and summed within- and among-lines to provide mean and variance for % exposed boulder for each 5 m increment and for the 100 m study reach. The PVC frame was itself divided into 6 subunits using rubber bands stretched across the frame to create a six-celled grid of "minor subunits". This system was created so that a die could be cast (ie. from one to six) in the field to randomly select a "major" or "minor" subunit for sampling substrate particles on the stream bottom. Based on random numbers generated by cast of the die, a "major sub-unit" was selected and a particle found within the "major subunit" was measured for longest length using a 50-cm tree caliper or 1.6 m measuring rod depending upon the size of the particle discovered. If several or many small particles were found in the "major subunit’ a second random number was generated to determine the location of the "minor subunit" to be sampled. Five random numbers were generated for each sq m cell across the stream cross-section and sampled in this manner. Mean particle size and variance were determined among sq m cells within-lines and among lines within-streams. Bedrock encountered on the stream bottom was assigned a value of 450 cm and silt particles too fine to be measured with the calipers were assigned a value of 0.01 cm. Large mean particle values with low variance therefore indicated bedrock bottoms and small mean particle values with low variance high silt environments.

For between-stream comparisons, attribute means were compared through analysis of variance (ANOVA)(GLM Procedure, SAS Institute, 1992). Depth and substrate measurements were normalized using a square root (x + 1) transformation. Slope, sinuosity, and chemical measurements were transformed using log (x + 1). Coefficient of variation (CV = Standard deviation/ mean) was used as a comparative measure of variability. Only untransformed means are reported and the use of the term "significant" indicates statistical significance (ie. P< 0.05) even though probability values are not specified in the report for simplification.

Standing Biomass of Algae and Invertebrates

At the beginning of the study, standing crops of algal and invertebrate biomass were estimated through benthic collections at approximately every 10 m increment using the previously described sq m cross-stream grid. In each sq m cell, estimates of particle sizes (described previously) were used to determine the predominate particle sizes found in the cell. Particles >30 cm in longest length (phi scale -8; Cummins 1962) were sampled using a standard Surber net (Surber 1937) and smaller particles through removal (see Kido 1996c for a description and rationale for the method). A minimum of three Surber scrapes and/or three rocks were sampled in each sq m cell depending on resident particle sizes. Substrate surfaces were scraped thoroughly with a stiff brush, the contents washed into a nitex fabric square and subsequently stored in 10 % buffered formalin. Surber nets sample an standard area of 0.09 sq m and rocks were measured for longest length and width to provide an estimate of rock surface area sampled. These values were used to calculate algal and invertebrate densities (biomass g/ sq m). In the laboratory, algae and invertebrates were processed, identified, and quantified as in Kido (1996a, 1996b). Algal samples were ashed after drying/weighing at 450^o C for three hours in a muffle furnace and results are reported as g AFDM per sq m (ash free dry matter). Invertebrate taxa were dried, and weighed directly; therefore, data are given as biomass g per sq m.

Stream Functional Processes

Hardboard multiplate samplers (Fig. 4) were used to assess the dynamics of algal/invertebrate colonization and deposition of sediment, coarse particulate organic matter (CPOM)(usually > 1mm pieces of vegetative matter), and fine particulate organic matter (FPOM) (usually organic detrital particules < 0.45 mm) in the study streams over the study period.

Figure 4. Hardboard samplers.

These samplers have a standard surface area of 0.16 m². Three samplers were used per replicate and three replicates were implanted in Alakahi and Onomea streams at low-middle-high elevation locations within the 100 m study reaches. Samplers within-replicates were secured to the stream bottom with lag-screws drilled into the rock substrate and placed within an area of about 1 sq m. Depth (m) and mean flow (cm/sec) for sampler replicates within-streams measured on the same day were 0.01 + 0.014 cm/sec (CV=3.00), 0.16 + 0.018 m (CV=0.344) and 0.05 + 0.016 cm/sec (CV=0.941), 0.14 + 0.024 m (CV=0.519) for Alakahi and Onomea streams respectively.

Samplers were left in the stream for 3-4 week periods. Upon removal, bottom plates from each sampler were stored in a dark container for total chlorophyll analysis and the remaining plates disassembled and scrubbed clean of organic material. This material was processed as described earlier for benthic samples. Algal, bryophyte and FPOM densities were determined by estimating % surface area coverage in the sample and multiplying this value by ash free dry weight (AFDW) of the entire sample. CPOM densities were estimated similarly or through direct removal of vegetative pieces with subsequent drying, weighing, and ashing. Sediment densities were assumed to be the inorganic residues remaining after organic material had been burned away during combustion in the muffle furnace. Samplers have identical surface areas; therefore, biomasses of algal and invertebrate taxa were only standardized by days of stream exposure giving data comparable as a rate of biomass accumulation over time (biomass g AFDM per day).

Total Primary Production

Total chlorophyll as an estimate of overall instream primary productivity was determined from the bottom plates of each sampler which were frozen after removal from the stream in dark containers until processing. Individual plates were removed from the freezer and scraped clean of organic material which were filtered under pressure on to a burnable filter (4.5 cm diameter). Filters were carefully placed into centrifuge tubes containing 10 mL methanol and returned to the freezer for 24 hr. In a dark room, 2.0 mL methanol and 0.4 mL of sample liquid were placed into a sample cuvette with an auto-pipette and mixed. Samples were read on a digital fluorometer (Sequoia-Turner, Model 450) zeroed on 2 mL of pure methanol. Filters were burned in a muffle furnace and weighed to determine AFDW. Plates were assumed to have identical surface areas; therefore, chlorophyll units were only standardized by days of stream exposure and are reported as accumulating chlorophyll units per day.

Results and Discussion

Physicochemical Comparisons

In the two study reaches compared, Onomea was found to be significantly deeper and wider than Alakahi with more uniform depth (ie. lower variance)(Table 1). This characteristic was likely maintained by supplemental flow diverted into Onomea from adjacent Kawainui stream entering above the study site. It was no surprise, therefore, that Onomea also had significantly higher flow velocity even though Alakahi had greater overall slope (Table 1). Higher stream velocity in Onomea was also supported by fewer changes in flow direction (ie. lower sinuosity) which allowed the stream to flow straighter in long, steep, high velocity channels (Fig. 5).

Figure 5. Sinuosity comparisons.

Alakahi, in comparison, meandered considerably exhibiting significantly greater sinuosity (Table 1) which resulted in shorter, slower-flowing channel units separated by pools and braided segments. Stream width in Alakahi, not surprisingly, also exhibited higher variance than Onomea due to marked widening of the stream between channel units (Table 1).

Table 1. Comparison of means and spatial variability (coefficient of variation) of physicochemical attributes of Alakahi and Onomea streams.

The two streams, although geographically adjacent, exhibited distinctly different substrata apparently because they cut through two distinct types of basalt lava formations. Indeed, it may likely have been differences in erosion resistance of these formations that ultimately dictated the stream’s physiognomies. Onomea flowed through a solid vein of erosion-resistant basalt bedrock which comprised most of the stream bottom. This bedrock substrate was found to be overlain by cobble-sized particles which constantly travel downstream. Deposition of this material occurred primarily in scour pools and low-slope segments. Onomea, therefore, exhibited large mean particle size (due to preponderance of bedrock), low mean % exposed boulder, but high variability in these attributes over the study reach (Table 1). Meandering, slower-flowing Alakahi cut through a less-erosion resistant basalt vein resulting in a talus-type substrata characterized by high coverages of meter-sized (or greater) exposed substrate and many areas of gravel/silt deposits (Table 1).

These physical habitat differences likely influenced measurable chemical differences in the streams themselves. Onomea, for example, was found to have significantly higher dissolved oxygen concentrations than Alakahi with higher variance suggesting that relatively high oxygenation was occurring in certain segments of the study reach.

These supersturated zones were determined to be areas of high velocity, slope and turbulence in Onomea not present in Alakahi. Mean water temperature and conductivity were significantly higher in Alakahi while no between-stream differences were determined for pH (Table 1). The Onomea study reach was more open in riparian canopy and thus more exposed to sunlight; however, lower stream temperatures may have been maintained by swifter flows and the influx of diverted water from Kawainui stream. As expected, higher conductivity was found in Alakahi as nearly totally closed riparian canopies likely inputed large quantities of organic detritus into the study stream reach.

These comparisons of Alakahi and Onomea suggested that these proximate streams differed significantly in physical characteristics and resultant dynamic physical processes occurring within-stream. The supplemented flow into Onomea from Kawainui likely played a significant role in influencing and maintaining these differences. The type of basaltic lava through which the streams flowed also had a major influence on the character of the stream bottom as well as on the manner in which the stream meandered. The question remained, at this point in the study, as to how these physicochemical differences between-streams influenced their biotic properties and processes.

Benthic Comparisons of Algal and Invertebrate Populations

Onomea was found to have greater species diversity and significantly higher biomass of algae and invertebrates than Alakahi based on initial density estimates obtained by sampling of the stream bottom (Table 2). In both Onomea and Alakahi, the blue-green alga Phormidium retzii was found in highest abundance forming dense, dark-green mats in thick layers of fine particulate organic matter (FPOM). An aquatic bryophyte (moss), tentatively identified as Champtochaete sp., was also found in both streams but in significantly higher densities in Alakahi. Immature midges (primarily adventive Cricotopus bicinctus)(Chironomidae) were the most abundant invertebrate found in both streams but in significantly higher densities in Onomea. Therefore while the two streams shared many of the same algal and invertebrate species, they almost invariably exhibited higher densities in Onomea (except for the moss).

Table 2. Algal and invertebrate species and density comparisons found in study reaches of Onomea and Alakahi streams based on benthic sampling (19-20 March 1997).

Native aquatic insects were only collected in Onomea. Adult male Megalagrion hawaiiense were observed and a few naiads collected in a small waterfall which emptied into the stream approximately 50 m upstream of the study site’s 100 M mark (~ 50 m elevation). Pupal cases with emerging native Neoscatella cilipes (Ephydridae:Shore flies) were collected during benthic sampling; however, no adults were captured while sweep-netting the stream. This suggests that populations of this active native stream fly in Onomea were small.

Functional Processes

Primary Production

Primary productivity as measured by accumulating total chlorophyll units per day was significantly higher in Onomea than Alakahi over the study period (Fig. 6).

Figure 6. Comparison of instream primary production.

This is supported by findings of greater initial standing crops of algal biomass discussed earlier for Onomea. Although Onomea was determined to be a more productive stream; both streams exhibited similar trends in primary productivity with a peak occurring in the spring, falling to lows in the summer, and increasing to moderate levels in the fall. In the absence of continuous flow and weather data, however, it is difficult to ascertain the respective roles played by natural variation of environmental factors such as periodic flood disturbance, light, and/or temperature in influencing the observed patterns in primary productivity.

Algal and Invertebrate Dynamics

Five filamentous algal and twenty-one invertebrate species were collected off hardboard multiplate samplers implanted in Onomea and Alakahi streams over the study period (Table 3). While no new algal species were collected as compared to those obtained during the initial benthic assessment, five previously uncollected arthropod taxa, a nemertean (proboscis worm), and at least two species of aquatic annelids were obtained from the implanted hardboard samplers. For nearly every algal and invertebrate taxa, rates of colonization were higher in Onomea as compared to Alakahi with the exception of an adventive dragonfly Ischnura sp. (not collected in Onomea) and aquatic Collembola (springtails)(Table 3). These data supported the notion that Onomea was a significantly more productive stream than neighboring Alakahi.

Table 3. Comparison colonization rates for algal and invertebrate species collected off hardboard multiplate samplers implanted in Onomea and Alakahi Stream (Apr to Oct 97).

For invertebrates, the alien midge Cricotopus bicinctus was determined to be the dominant species collected in terms of accumulating biomass (Table 3). This is typical of Hawaii’s streams today; however, the scarcity of alien caddisflies (Cheumatopsyche petti and Hydroptila arctia) was rather atypical as they are generally expected to be more common in flowing streams of this nature. Of particular interest, was the relative dominance of micro-crustacea and soft-bodied worm-like species collected in both Onomea and Alakahi through use of implanted samplers (Table 3). These forms did not show up in the initial assessment which utilized standard stream-bottom sampling methodologies. No identification keys were available for these tiny forms and little is known about their ecology in Hawaii; however, their presence in lotic habitat and similarities in in situ colonization rates suggest that they are resident species in these two Hamakua Coast streams. Overall rates of algal colonization were consistently higher in Onomea as compared to Alakahi regardless of season (Fig. 7, 8).

Although the same algal species were found in both streams, their colonization dynamics differed markedly between-stream. For example, in Onomea stream two algal species, the blue-green Phormidium retzii and the green Spirogyra sp., consistently dominated the samplers with higher rates of colonization than other species (Fig. 7). A significant bloom occurred during fall (Oct 97) with both species exhibiting extremely rapid colonization as rates for other species declined (Fig. 7).

In contrast, samplers in Alakahi were dominated by three algal species. P. retzii, Cladophora sp., and Stigeclonium subsecundum (Fig. 8). Spirogyra sp., one of the two dominant Onomea algae, was relatively scarce in Alakahi and not a very productive species (Fig. 8) in that stream. The fall algal bloom observed in Onomea also occurred in Alakahi but only involved a single species (Cladophora sp.). While other resident algae in Alakahi declined in abundance, Cladophora sp. achieved comparable rates of growth to algal species in Onomea (Fig. 7,8).

It is difficult to interpret these findings given that few cross-stream comparisons have been attempted in Hawaii, particularly those having to do with instream primary productivity and algal succession. It seems clear that the differences in physical characteristics determined for the two streams were translated into biotic differences as well. The importance of physical habitat (especially current regime) in influencing algal succession is well established (Stevenson et al. 1996). For algae, species abundances were likely mediated and maintained by relative differences in incident light, flow velocity, and nutrient input. Flood disturbance occurring unpredictably during the study period likely also played an overriding role in guiding the patterns observed as the ability for flood events to ‘reset’ levels of algal abundance in Hawaiian streams has been previously demonstrated (Kido, 1997a). For the sake of generalizations, Onomea can best be described as a more-productive, swift-current stream and Alakahi as a less-productive, slow-current stream. Each system functions within a set of dynamic processes dictated by individual physionomies.

For invertebrate species, higher algal productivity in Onomea resulted in higher abundances (colonization rates) of the alien midge, Cricotopus bicinctus. This swift-water chironomid is becoming increasingly abundant in Hawaii’s streams particularly those supporting robust populations of algae upon which they feed and reproduce. Interestingly, the abundances of micro-arthropod/worm-like species were similar in the two streams suggesting that these invertebrates were less dependent on algal abundance (ie primary productivity). Populations of these species may be more dependent upon terrestrial (external) sources of energy from organic matter input into the streams (ie. coarse and fine particulate organic matter).

Organic Matter Inputs and Sediment Dynamics

Onomea and Alakahi streams were found to have similar basal rates of CPOM, FPOM, and fine sediment accumulation over the study period; however, Onomea exhibited significantly greater peaks in deposition of these materials (Fig. 9).

Figure 9. Organic matter inputs and sediment dynamics in Onomea and Alakahi.

The similarities in pattern of fine sediment and FPOM accumulation in the two streams suggested that the dynamics of movement of these materials were being driven by processes occurring at the watershed scale (eg. heavy rain and/or flood events). The significantly greater rates of fine sediment and FPOM deposition which periodically occurred in Onomea (Fig. 9) may have been related to events occurring in the Kawainui stream watershed via diverted water inputs and/or land-based processes/activities specific to the Onomea watershed. Similar causes may be invoked to explain spikes in CPOM deposition in Onomea which consisted primarily of decomposed leaf and branch fragments. In Alakahi, the large late summer-fall (Oct 97) input of CPOM (Fig. 9) consisted primarily of palm fruits/seeds and fresh leaves/branches which indicated a more local source of this material from the extensive riparian canopy which bordered the stream channel along the study reach. These findings suggested that coarse vegetative material in the study reaches entered the streams from areas upstream as well as directly in closed riparian canopy locations. The organic material accumulating in the study sites from upstream were highly decomposed indicating that breakdown occurs in situ as the material moves downstream. The coarse material provided a steady source of FPOM for reaches further downstream. Basal stream flow conditions moved FPOM, CPOM, and fine sediment steadily into the study reaches with occasional large pulses of material being deposited during flood events. In the absence of comparative data from other streams, it is not possible to determine if the rates of deposition are higher or lower than other streams in the State. Nearly all of the riparian vegetation along the segments of Onomea and Alakahi for this study were non-native, the upper reaches of which ran through abandoned sugarcane fields. This situation is likely typical of Big Island Hamakua streams today. Judging from the results of this study, watersheds/riparian zones supply significant quantities of organic material locally to these streams; however, little is known about the fate of this material or its role (or affect) in/on stream function.

Conclusions

The stream reaches studied in Onomea and neighboring Alakahi were found to differ significantly in physical attributes of depth, width, flow, slope, sinuosity, dissolved oxygen, and substrata even though sites were located at nearly identical locations. These physical differences translated into biotic differences as well with Onomea exhibiting significantly greater primary and secondary productivity. Native insect species were only found in Onomea; however, their populations were not robust. Interestingly, while the two streams shared similar assemblages of algae and invertebrates, population dynamics within-stream differed markedly yet exhibited similar overall seasonal trends. Coarse/ fine particulate organic matter and fine sediment continually settled in the study reaches coming from sources upstream or directly from localized riparian areas. Very large periodic inputs of this material were determined in Onomea and to lesser degrees in Alakahi. Diverted water from Kawainui stream into Onomea likely plays an important role in governing physical and functional processes in reaches below the point of entry of diverted water.

As for providing habitat for native amphidromous macrofauna (ie. ‘o’opu, hihiwai, and opae), the study reach in Onomea probably would be predicted as supporting higher populations based on the between-stream comparisons. Onomea’s swifter-flow, higher oxygenation, and more complex substrate/channel unit characteristics provide more suitable physical habitat. The same conclusion would be reached based upon the availabilities and abundances of food species; however, overall densities of native macrofauna in both streams would be predictably low. Aquatic insect food biomass was quite low in both streams and primary production was based primarily on the cyanophyte Phormidium retzii with periodic blooms of Spirogyra sp. and Cladophora sp. In this environment, ‘o’opu-nakea (Awaous guamensis) would probably be present (but not abundant) while hihiwai (Neritina granosa) and opae (Atyoida bisulcata) would be predictably rare most of the time because of the rarity of chlorophytes. The ready availability of epiphytic and free-living diatoms in Onomea (and to a lesser extent in Alakahi) indicated that the obligate herbivore, Sicyopterus stimpsoni (‘o’opu-nopili), would at least be a common visitor to the study reach if not resident at times. Similarly, the ‘o’opu-alamo’o (Lentipes concolor) should be present, particularly with the micro-crustacean fauna available as prey. These food species appeared to fluctuate markedly in abundance during the study period and these fluctuations may be a factor which limits native macrofaunal populations over the long-term in both streams.

Acknowledgements

We thank Dan Lutkenhouse and the Hawaii Tropical Botanical Garden (HTBG) for funding this study. We also appreciate greatly the assistance of Scott Lucas and the HTBG staff in facilitating the logistics of the field work. We also thank Bob Nisihimoto (State Aquatic Resources Division) and Jeff Burgett (U.S. Fish and Wildlife Service) for their welcomed advice in the planning stages of the study. Appreciation is also extended to Bob Sheath (University of Guelph) for identifying algae and to Gordon Smith (State Department of Health) for evaluating stream bioassessment protocols in Onomea and Alakahi. We are also extremely grateful to Harvey Sr. and Phyllis Chong for their kind hospitality in Hilo.

Literature cited

Cummins, K.W. 1962. An evaluation of some techniques for the collection and analysis of benthic samples with special emphasis on lotic waters. Am. Mid. Naturalist 67(2):477-504.

Hawaii Stream Assessment. 1990. A preliminary appraisal of Hawaii's Stream Resources, Report R84. Prepared for the State Commission on Water Resource Management by the Hawaii Cooperative Park Service Unit (National Park Service), December 1990, Honolulu.

Kelly, J.R. and S.A. Levin. 1986. A comparison of aquatic and terrestrial nutrient cycling and production processes in natural ecosystems, with reference to ecological concepts of relevance to some waste disposal issues. In: G. Kullenberg (ed.), The Role of the Ocean as a Waste Disposal Option, p. 165-203. Reidel Publishing, Boston.

Kido, M.H., D.E. Heacock, and A.M. Brasher. 1994. Comparison of visual estimation methods applied in Hawaiian streams. In: Proceedings of the International Symposium on Hawaiian Stream Ecology, Preservation and Management, 1993 Hilo, Hawaii, C.M. Way (ed.). U.S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg, MS.

Kido, M.H. 1996a. Diet and food selection in the endemic Hawaiian amphidromous goby, Sicyopterus stimpsoni (Pisces: Gobiidae). Environ. Biol. Fish. 45(2):199-209.

1996b. Morphological variation in feeding traits of native Hawaiian stream fishes. Pac.Sci. 50(2):184-193

1996c. Recovery Processes in Hawaiian Streams. In: Proceedings of a Symposium on Stream Restoration in the Hawaiian Islands, 1994 Honolulu, Hawaii, W.Devick (ed.), State Department of Land and Natural Resources, Aquatic Resources Division.

1997a. Food relations and coexistence between native Hawaiian stream fishes. Env.Biol.Fish. 49:481-494.

Palmer, M.A., C.C. Hakenkamp, and K. Nelson-Baker. 1997. Ecological heterogeneity I streams: why variance matters. J. N. Am. Benthol. Soc. 16(1): 189-202.

Peterson, C.G. 1996. Response of benthic algal communities to natural physical disturbance. In: R.J. Stevenson, M.L. Bothwell, and R.L. Lowe (eds). Algal Ecology. Academic Press, San Diego, CA. Pp. 375-402.

Surber, E. W. 1937. Rainbow trout and bottom fauna production in one mile of stream. Trans. Am. Fish. Soc. 66:193-202.

Timbol, A.S. and J.A. Maciolek. 1978. Stream channel modification in Hawaii: PartA: Statewide inventory of streams; Habitat factors and associated biota. FWS/OBS - 78/16, April 1978.

Stream Monitoring	GIS	Stream Surveys	Aquatic Organisms	Conference Papers	Islands Survey