Tree diversity in sub-montane and lower montane primary rain forests in Central Sulawesi

The tree diversity of sub-montane and lower montane primary forests is studied in plot-based inventories on two sites in Lore Lindu National Park, Central Sulawesi. Out of 166 species in total, 50 % are new records for Sulawesi (19 %) or the Central Sulawesi province (31 %). Species richness decreases with altitude. In the submontane forest, the highest Family Importance Values (FIV) are reached by the Lauraceae, Fagaceae, Sapotaceae, Moraceae and Euphorbiaceae. In the lower montane forest, the Fagaceae are of major importance (FIV 71.9), followed at some distance by the Myrtaceae, Elaeocarpaceae and Lauraceae. For each site, a group of important families is identified that is of minor importance or absent on the other site. The comparison of basal area (BA), number of species and FIV with published plot-based studies in sub-montane and lower montane primary forests in Malesia (Borneo, Sulawesi, Papua New Guinea) reveals: 1) with 35.4 and 37.1 m² ha-1, the BA is comparable to that measured in Borneo and Papua New Guinea, but does not support previous findings of extremely high BA in Sulawesi forests; 2) species richness is comparable to that in Borneo and other Sulawesi forests, but lower than in Papua New Guinea; 3) decrease in diversity with altitude is in accordance with findings in Borneo; 4) in sub-montane forests, the Lauraceae are generally important; the Sulawesi studies are closely related to those from Papua New Guinea; 5) the lower montane forests have the Fagaceae and Myrtaceae as most important families in common.


RESEARCH ARTICLE InTRoduCTIon
A survey of plant species diversity and endemism across five major Malesian islands has indicated that the island of Sulawesi is intermediate for these measures (Roos et al. 2004).This is remarkable, because one would expect at least high endemism rates due to the isolation of the Wallacean island from the Sunda Shelf during the Quaternary period (Primack & Corlett 2006).Cannon et al. (2007) pointed out that this mediocrity might be related to the fact that collection rates on the island are among the lowest in Indonesia and to the limited taxonomic study.Hence, additional collections, especially from remote and primary forest areas in Sulawesi, are needed to further our knowledge of the island's plant diversity.Our plot-based tree inventories in primary forests of Lore Lindu National Park, Central Sulawesi, explore the species diversity of sub-montane and lower montane forests.We present a large number of new taxonomic records for Sulawesi or the Central Sulawesi province compared to the Checklist of woody plants of Sulawesi (Keßler et al. 2002).
Especially in montane forests of the Malesian tropics, detailed tree surveys are limited.Quantitative altitudinal transect studies are known from Mt Kinabalu, Borneo (Kitayama 1992, Aiba & Kitayama 1999, Aiba et al. 2005).A tree diversity study on a one-hectare-plot was carried out by Wright et al. (1997) at 900 m altitude in Papua New Guinea.In Lore Lindu National Park, Central Sulawesi, primary forests were studied at 1 100-1 200 m altitude by Kessler et al. (2005).The present study is the first to deal with the primary forest of Sulawesi at different elevations.We aim at identifying the most important tree families in our surveyed forests as well as detecting changes in tree family composition between sub-montane and lower montane primary forest sites.

Study area
The two primary forest study sites are located in Lore Lindu National Park, Central Sulawesi, Indonesia.Most parts of the protected area are covered by upland and montane forests on intermediate soils.The forest condition is good to old growth (Cannon et al. 2007).
The first site is situated in Pono Valley at 1 050 m altitude (S 01°29.6',E 120°03.4',GC-WGS 84).The habitat is a submontane old growth forest on Ferralsol (FAO 2006) developed on metamorphic rocks in a stable level terrain on a mid-slope.Pono Valley is one of the Sulawesi Throughfall Displacement Experiment test sites hosted by the collaborative research centre SFB 552 (University of Göttingen).
The second site is located in the Bariri Forest at 1 400 m altitude (S 01°39.5',E 120°10.4',GC-WGS 84).The habitat is a lower montane old growth forest on Nitisol (FAO 2006) developed on sedimentary substrate on a level plateau.A 70 m tall meteorological scaffold tower is constructed in the centre of the forest.The forest shows small-scale disturbances close-by the tower caused by the tower construction.

Field sampling
Plot-based tree inventories were carried out in the period from July to September 2006.Collections were completed in 2007.
Plot size was 40 × 60 m (0.24 ha) divided up into a 10 × 10 m grid.All trees of diameter at breast height (dbh) ≥ 10 cm were surveyed.Within the 10 × 10 m grid, 5 × 5 m-sized subplots were nested (0.06 ha).Therein, understorey trees of dbh 2-9.9 cm were additionally sampled.All trees were permanently tagged, pre-identified, their structural parameters (dbh, total height, trunk height) and their spatial position recorded.In Pono Valley six plots were installed ± parallel to each other aligned to the Sulawesi Throughfall Displacement Experiment test plots (total area 1.44 ha).In the Bariri Forest, three plots were set up in a radial arrangement around the meteorological tower avoiding disturbed areas (total area 0.72 ha).

Tree species identification
Tree species identification was based upon about 2 000 specimens (collection numbers HC) collected from tagged trees and supplementary trees in flower or fruit.Specimens were deposited at CEB, GOET, K and L.
Tree species were identified by H. Culmsee using the collection at the National Herbarium of the Netherlands, University of Leiden branch, as reference and by specialists for Elaeocarpaceae (M.J.E.Coode, K), Moraceae (C.C.Berg, L) and Myristicaceae (W.J.J.O.de Wilde, L).Taxa difficult to identify to species, especially in the Myrtaceae, were distinguished as separate species based on morphology of vegetative characters (leaves, twigs and barks).

Tree diversity analysis
Species-level presence/absence data related to study site included all tagged trees of dbh ≥ 10 cm and were complemented by supplementary species found in the understorey subplots.The assessments as new records for Sulawesi or Central Sulawesi were based on comparison with the Checklist of woody plants of Sulawesi (Keßler et al. 2002).
As the size of the sampled area varied between sites, samplebased rarefaction curves (Gotelli & Colwell 2001) were calculated using EstimateS v8.0.0 (Colwell 2006) to assess the comparability of species richness per site.Sample units were based on individual counts within the 10 × 10 m grid, i.e. 24 samples correspond to one plot (0.24 ha).The analysis included all trees surveyed (dbh 2 -9.9 cm and dbh ≥ 10 cm).
On the family level, relative frequency (based on the enumeration of individuals) and basal area (based on the dbh measured) were calculated.The Family Importance Value (FIV, Mori et al. 1983) was used to assess the contribution of each family to the stand.FIV combines relative family richness (number of species), relative density (number of individuals) and relative dominance (basal area) into one value.Between-site comparison on the family level took into account trees dbh ≥ 10 cm.

Species richness
The sub-montane forest in Pono Valley (alt. 1 050 m) accommodates 123 tree taxa assigned to 42 families.The lower montane Bariri Forest (alt. 1 400 m) is much species poorer with 74 tree taxa out of 36 families (Table 1).
Although the size of the sampled area in the Bariri Forest (0.72 ha) is only half of that in the Pono Valley (1.44 ha), the  difference in species richness is not an artefact.The samplebased rarefaction curve (Fig. 1) shows that already at half of the effectively sampled area in the Bariri Forest the curve starts rising slowly.If the curve is extrapolated, it reaches ± 85 taxa at 1.44 ha.In contrast, the number of species is expected to exceed 100 already at 0.72 ha in Pono Valley.
The combined Pono Valley and Bariri Forest species list (Table 2) comprises a total of 166 taxa assigned to 49 families.Less than 1 % of the sampled trees remained unidentified.At Pono Valley, the trees of dbh ≥ 10 cm include 104 taxa with an additional 19 supplementary species found among the trees of dbh 2-9.9 cm.The Bariri Forest comprises 60 tree species of dbh ≥ 10 cm with 14 taxa additionally found among the trees of dbh 2-9.9 cm.The two sites have 33 species in common.The highest number of species is found in Lauraceae (19 spp.), Euphorbiaceae (13 spp.), Elaeocarpaceae (12 spp.), Meliaceae (12 spp.), Rubiaceae (10 spp.), Moraceae (9 spp.) and Myrtaceae (9 spp.).
A total of 50 % are found to be new records, with 19 % as new records for Sulawesi (32 spp.) and 31 % as new to Central Sulawesi (51 spp.).The new records for Sulawesi include on family level the Gesneriaceae and the Hamamelidaceae.Tree ferns are explicitly not included in the checklist (Keßler et al. 2002), but are included in the present list.One new species was described based on the collections from Pono Valley (Culmsee 2008).

Community composition
The sub-montane and the lower montane forests have ten important tree families in common (based on FIV, Table 3).In Pono Valley, the Lauraceae have the highest importance (FIV = 30.2),closely followed by the Fagaceae, Sapotaceae, Moraceae and Euphorbiaceae.In the Bariri Forest, the Fagaceae are by far the most dominant family (FIV = 71.9).They are followed by the Myrtaceae, Elaeocarpaceae and Lauraceae.
The families Juglandaceae, Oleaceae and Theaceae have high importance in the lower montane forest, but their importance is low in the sub-montane forest.In contrast, the Asteraceae, Meliaceae, Myristicaceae, Rubiaceae and Rutaceae are important families in the sub-montane forest, but their importance decreases considerably in the lower montane forest.The Annonaceae, Cyatheaceae and Dracaenaceae are present in the sub-montane forest, but they are absent in the lower montane forest.

dISCuSSIon
The large number of new records of tree species for the island of Sulawesi or its Central province is remarkable, in particular as the Lore Lindu National Park is among the parts of Sulawesi for which the largest plant collections are available (Cannon et al. 2007).For the whole island, the plant collection density is one of the lowest in Malesia (Roos et al. 2004), with 25 per 100 km².Furthermore, only few plant collections are available from the quite extensive forests in good or old-growth conditions in other parts of Sulawesi (Cannon et al. 2007).That leads to the assumption that the relative number of new records or even new species could potentially be even higher in primary forests of other, less investigated parts of Sulawesi.
Compared to plot-based studies at similar altitudes in Malesia (Table 4), the plots in the present study are less species   3 Family Importance Value (FIV; R: within-site ranking of families sorted by FIV), number of species (# sp), relative frequency in % (Rel FQ) and basal area (BA ha -1 ) for the most important tree families (FIV ≥ 5.0) at Pono Valley (alt. 1 050 m) and Bariri Forest (alt. 1 400 m) based on trees dbh ≥ 10cm.

Table 1
Basic Pono Valley and Bariri Forest tree inventory metrics.