Nodule Biomass of the Nitrogen-Fixing Alien Myricafaya Ait. in Hawaii Volcanoes National Park!

Nodule Biomass of the Nitrogen-Fixing Alien Myrica faya Ait. In Hawaii Volcanoes National Park!

D OUGLAS R. TURNER AND PETER M. VITOUSEK 2

ABSTRACT: Myricafaya forms a nitrogen-fixing symbiosis in which fixation takes place in specialized root nodules. The biomass of these nodules was greater in open-grown than shaded individuals ofM yricafaya, and was greater in large than small individuals. All Myricafaya examined, including seedlings and those growing epiphytically, had active nodules. Nitrogen fixation by invading Myrica faya could alter patterns of primary succession in Hawaii Volcanoes National Park.

Myrica faya AlT. IS AN AGGRESSIVE INVADER volcanic ash, and inthe open in volcanic ash of most of the major Hawaiian Islands depo sits as young as 10- 12 years . (Whiteaker and Gardner 1985). It grows rap The nitrogen-fixing potential ofthe Myrica/ idly, forms dense, nearly monospecific stands, Frankia symbiosis is of particular concern in and can reach 20 m in height (Smathers and HAva because young volcanic soils are very Gardner 1979). Myricafaya represents a par low in available nitrogen (Vitousek et al. 1983, ticularly serious problem because it forms a Balakrishnan and Mueller-Dombois 1983). symbiotic association with the nitrogen-fixing Inva sion by Myrica faya could supplement actinomycete Frankia (Miguel and Rodri the slow accumulation of nitrogen from rain gues-Barrueco 1974, Mian et al. 1976) and fall and nonsymbiotic fixation with more rap hence can use nitrogen from the atmosphere id symbiotic fixation. Since susceptibi lity to rather than the soil. This symbiosis , unlike the invasion of Hawaiian ecosystems is dependent better known legume-Rhizobium association, in part on soil fertility (Gerrish and Mueller apparently was absent from the native Hawai Dombois 1980), Myrica faya could therefore ian flora. M yricafaya is considered one ofthe affect both the productivity and the possibility worst weeds in Hawai'i (Smith 1985) because of further invasions of HAva ecosystems. of its ability to fix nitrogen and its potential to We collected descriptive information on the form monospecific stands. presence and distribution of nitrogen-fixing Myricafaya invaded the Hawaii Volcanoes root nodules on Myrica faya in HAva and National Park (HAVO) in 1961 (Doty and quantitative information on nodule biomass. Mueller-Dombois 1966). By 1978, its range Myricafaya root systems were harvested from was 600 ha despite strenuous control efforts . plants in three areas-open, partially shaded, Shortly thereafter, park-wide control was and under a Metrosideros-Cibotium canopy. abandoned, and Myrica faya had occupied The results of the harvests were used to de 12,000 ha by 1985 and was still expanding velop regressions relating nodule biomass to (Whiteaker and Gardner 1985). Populations basal area of individuals of Myrica faya. of M yrica faya are densest in open seasonal montane forests and pastures, but they are found in the understory of closed Metroside ros polymorpha forests, in stands damaged by STUDY SITES All study sites were at 1150- 1250 m eleva tion in HAVO, near the summit of Kilauea Volcano on the island of Hawai'i. Mean an I Manuscript accepted February 1987. 2 Departm ent of Biological Sciences, Stanford Univer nua l temperature is 17°C, and mean annual sity, Stanford , California 94305. precipitation near 2000 mm. 186 Nodule Biomass of Myricafaya-TURNER AND VITOUSEK 187

The open site (site # 1) is located in Devas at 70cC for 72 hours and then weighed. Ap tation Area, a deep deposit of volcanic cinder proximately 100 person-hours were required resulting from a 1959 eruption of Kilauea Iki to harvest and sort the nodules from each of Crater. The site is now sparsely colonized by the largest Myrica faya sampled. Metrosideros polymorpha, Myrica faya, and several other native and exotic woody plants; community composition has been sampled RESULTS AND DISCUSSION repeatedly since the 1960's (Smathers and Mueller-Dombois 1974, Wright 1985). We observed M yrica faya rooting in bare Adjacent to Devastation Area, much ofthe volcanic cinder and in organic soils, establish Metrosideros-dominated forest survived shal ing in cracks in old pahoehoe lava with a thin lower deposits «2 m) ofcinder in 1959. This veneer of ash, and perched on decaying trunks area (site #2) is now an open-canopied Me or epiphytically on the tree fern Cibotium trosideros forest , with rapid colonization of cham issoi. Nodules were present on all of the Myrica faya occurring under and at the edge individuals examined, including seedlings < 2 of M etrosideros crowns. Areas of bare cinder em tall and epiphytes on Cibotium (in which remain , and most of the colonizing Myrica nodules occurred within the outer fibrous faya receive some full sunlight each day. trunk of the tree fern). We found nodules up The third site (site # 3) is located within to 13m away from the base ofa tree, and up to a MetrosiderosjCibotium forest developed on 2.2 m deep. Masses ofbranched nodules up to a 1790 ash deposit near Thurston lava tube . 15 x 8 em in size were found, although most Myricafaya is present as scattered individuals clusters ranged from <.5 x .5 em to 3 x 3 in the forest understory; occasional larger in em. All of the nodules tested actively reduced dividuals reach into the subcanopy. acetylene to ethylene, indicating the presence of the nitrogen-fixing nitrogenase enzyme system. The aboveground growth form of Myrica METHODS faya varied among the sites. Individuals from Myrica faya growth form , root systems, the forest understory (site 3) were taller in and nodules were examined in all of the sites proportion to their diameter (Figure 1) and to develop an understanding ofranges ofvari less branched than open-grown individuals. ation. Six or seven individuals were then se Below-ground growth forms also differed; lected to represent a size range in each site, roots in the forest understory were noticeably and diameter and height (of each stem for sparser and appeared randomly distributed. multiple-stemmed individuals) were recorded. In the open and partially shaded sites, Myrica Root systems were then excavated using trow faya roots grew and proliferated more actively els and spoons; digging usually progressed in the organic-rich soil under plant canopies from the basal area outward, following roots than in areas ofbare cinder. Many roots grew to their branched and numerous tips. Clusters into the organic mat surrounding the trunk of of nodules were removed as they were en adjacent individuals ofMetrosideros polymor countered and stored in polyethylene bags. pha, and some were appressed against the When a complete root had been unearthed, Metrosideros trunk. One or two roots of the it was removed to allow access to deeper two largest open-grown individuals reached roots. Nodules were washed in the laboratory, the previous soil surface, which was buried sorted carefully to remove roots and debris, under more than 2 m of 1959 cinder. and washed again. Functional nodules were Masses of nodules on different-sized indi identified by their light-brown to brown color viduals are summarized in Figure 2. For a and the presence of vascular connections to given basal area, nodule biomass was greater roots; dark, dry or partially decomposed nod in the open-grown (site 1) than in the partially ules did not reduce acetylene effectively and shaded individuals (site 2), which in turn was were discarded. The active nodules were dried greater than under closed canopies (site 3). 188 PACIFIC SCIENCE, Volume 41,1987

-rn C Ol ~ -eQ)

20 40 60 80 100 120

BASAL AREA (cm2)

FIGURE I. Diame ter versus height of Myrica faya stems grown in the open ("I"), und er partial shade ("2 "), and under a closed forest canopy (" )").

The relationship between nodule mass and The proportion of variation explained by the basal area was asymptotic in sites 1 and 2 but regressions was .99, .94, and .99 in sites 1,2, linear in site 3. Non-linear regressions of the and 3 respectively. form The association between canopy cover and nodule biomass reported here (Figure 2) has been observed in other studies of Frankia nodulated plants (Binkley 1981, Bormann and Gordon 1984). This association could be where x is basal area in ern", B1 and Bz coeffi caused by greater energy availability for N, cients, A the asymptote, and Y nodule biomass fixation in the open -grown plants, or alter in grams, were fit to the results from site I and natively by slower growth and hence lower 2; linear regression was used in site 3 (slope = demand for fixed nitrogen in shaded plants. .33). B1 and Bz for site I were 27.5 and 0.18 , We did not anticipate the strongly asymp and for site 2 were 58.4 and 0.16 respectively. totic nature of the relationship between basal Nodule Biomass of Myricafaya- TURNER AND VITOUSEK 189

120

100

-~ 80 -CJ en lJJ :5 60 oCJ z

o o 20 40 60 80 100 120

BASAL AREA (cm2)

FIGURE 2. Basal area of Myricaf aya (em") versus nodule biomass in grams. Lines are least-squares regressions (see text). Trees were sampled in the open ("I"), under partial shade (" 2"), and under closed forest canopy ("3 ").

area and nod ule biomass in sites 1 and 2. We Overall, the information summarized here see three possible causes for the relatively demonstrates that invasion by Myrica faya small nodule mass in the largest individuals. could represent a substantial perturbation to First, Smathers and Gardner (1979) identified primary succession in Hawaii Volcanoes Na an inverse correlation between stem diameter tional Park. If Myrica fa ya nodules fix only and vigor of Myrica faya . Second , the avail 1 ,liM Nz/g/hr (a low rate for the Frankia ability of fixed nitrogen is elevated in the soil symbiosis), then a single partially-shaded in beneath large individuals of M yrica faya dividual could fix 14 gjyr . Populations of (P. A. Matson, personal communication), so Myricafaya in a partially-shaded site adjacent perhaps they require that less nitrogen be pro to the Devastation Area are approximately vided by energetically costly nitrogen fixation . 1100 individuals per hectare (Whiteaker, per Finally, the large individuals may have slower sonal communication), representing potential relative growth rates, and hence relatively nitrogen fixation of 15 kg/ha /yr. This addi lower demands for nitrogen. tional nitrogen represents a substantial modifi- 190 PACIFIC SCIENCE, Volume 41, 1987

cation to primary succession on these nitrogen two tropical rainforests on Oahu, Hawaiian deficient young volcanic soils. Islands. Phytocoenologia 8: 237-295. MIAN, S., A. BOND, and C. RODRIGUES BARRUECO. 1976. Effective and ineffective root nodules in Myricafaya Ait. Proc. Royal ACKNOWLEDGMENTS Society of London B 194:285-293. We thank P. A. Matson, L. R. Walker, D. MIGUEL, c., and C. RODRIGUEZ-BARRUECO. Wang, and L. Whiteaker for assistance in the 1974, Acetylene reduction activity of de field and laboratory, and D. Mueller-Dombois, tached root nodules of Myricafaya Ait. Plant C. P. Stone, L. R. Walker, and L. Whiteaker and Soil 41 :521-526. for constructive comments on the manuscript. SMATHERS, G. A., and D. E. GARDNER. 1979. C. P. Stone and his research staff at HAVO Stand analysis of an invading firetree (My greatly facilitated our efforts in the Park. Re rica faya Aiton) population, Hawaii. Pac. search was supported by NSF grant BSR-84 Sci. 33: 239-255. 15821 to Stanford University. SMATHERS, G. A., and D. MUELLER-DoMBOIS. 1974. Invasion and recovery of vegetation after a volcanic eruption in Hawaii. Na tional Park Service Science Monograph 5, LITERATURE CITED 129 p. BALAKRISHNAN, N. , and D. MUELLER SMITH, C. W. 1985. Impact of alien plants on DOMBOIS. 1983. Nutrient studies in relation Hawai'i's native biota, p. 180-250. In Stone, to habital-types and canopy dieback in the C. P., and J. M . Scott (eds), Hawaii's terres montane rain forest, island of'Hawai'i. Pac. trial ecosystems: preservation and manage Sci. 37: 339-359. ment. Cooperative Park Studies Unit, Univ. BINKLEY, D. 1981. Nodule biomass and acety of Hawaii, Honolulu. lene reduction rates of red alder and Sitka VITOUSEK, P. M., K . VAN CLEVE, N. BALAKRI alder on Vancouver Island, British Colum SHNAN, and D. MUELLER-DoMBOIS. 1983. bia . Canadian J. of Forest Research II: Soil development and nitrogen turnover in 281-286. montane rain forest soils in Hawai'i. Bio BORMANN, B. T., and J. C. GORDON. 1984. tropica 15: 268-274. Stand density effects in young red alder WHITEAKER, L., and D. E. GARDNER. 1985. plantations: Productivity, photosynthate The distribution of Myricafaya Ait. in the partitioning, and nitrogen fixation. Ecology State of'Hawai'i. Cooperative National Park 65:394-402. Resources Study Unit, Hawai'i, Technical DoTY, M. S., and D. MUELLER-DoMBOIS. 1966. Report 55. Department ofBotany, Univ. of Atlas for bioecology studies in Hawaii Vol Hawaii, Honolulu. 31 p. canoes National Park. Hawaii Agricultural WRIGHT, R. A. 1985. Shrub population dy Experiment Station Misc. Pub. 89. namics and succession on volcanic cinder GERRISH, G., and D. MUELLER-DoMBOIS. 1980. in Hawaii. MS Thesis, Univ. of Hawaii, Behavior of native and non-native plants in Honolulu. 369 p.