Relationships between soil chemical properties and forest structure, productivity and floristic diversity along an altitudinal transect of moist tropical forest in Amazonia, Ecuador.

Abstract

In a transect from mature upper lowland to montane forests in NE Ecuador, in the area of the Sumaco Biosphere Reserve (SBR), we tested the hypotheses that (1) the availability of P is low in low-elevation forests, but increases upslope, while the availability of N is relatively high at low elevations but decreases with elevation, and (2) increasing amounts of calcium, magnesium and potassium are stored on top of the soil with progressive humus accumulation toward higher elevations, likely to improve nutrient availability. In each 20 plots at 500, 1000, 1500 and 2000 m a.s.l., we measured about 20 soil chemical parameters in undisturbed natural forests (80 plots in total) including in situ N net mineralization and nitrification rate (NNM and NNI, buried bag method), plant-available phosphorus (Pa, resin-bag method), and salt-exchangeable calcium, potassium and magnesium concentrations (Caex, Kex, Mgex) in the organic and mineral topsoils. Pa showed a large increase with elevation, supporting the first hypothesis. While mass-related NNM and NNI rates and also organic matter C/N ratio in the topsoil remained invariant along the slope, NNM and NNI rates per ground area decreased by about 40% from 500 to 2000 m due to a lowered bulk density of the topsoil at higher elevations. The organic layers play a key role not only for N and P supply, but also as storage of exchangeable Ca, Mg and K. Caex, Mgex and Kex increased with elevation per ground area and also per humus mass, supporting the second hypothesis. The dependence of aboveground biomass and productivity on soil fertility in tropical forests is not fully understood since previous studies yielded contrasting results. Here we quantify aboveground biomass (AGB) and wood productivity, and the impact of soil chemistry on these parameters. We measured more than 20 soil chemical parameters and inventoried all trees ≥ 10 cm dbh and determined stem diameter growth with dendrometer tapes in 32 plots. Tree basal area reached the highest values at 1500 and 2000 m, whe reas AGB and productivity measures did not vary with elevation; across all plots, basal area averaged 47.2±1.9 m2 ha-1, AGB 336±17 Mg ha-1, and coarse wood productivity 1.6 ±0.2 % yr-1 (means ± SE). Tree coarse wood growth co-varied strongest with plant-available P and exchangeable Mg in the soil organic layers. These two elements most likely control tree productivity at both lowland and lower montane elevations. Stand basal area and AGB, which are influenced by both productivity and tree longevity, co-varied primarily with the exchangeable concentrations of Mg, Ca and K in the organic layers. The availability of N had a surprisingly small influence on forest biomass and productivity across the studied forest types, thereby contradicting the hypothesis of increasing N limitation of tree growth with elevation in tropical mountains. Our results indicate that wood biomass and productivity of equatorial Andean forests are influenced by more than one nutrient species, and that N does not play a key role. We tested the hypotheses that, with elevation, leaf area index (LAI) decreases, and diffuse transmittance of photosynthetically active radiation (DT) increases, and that tree species diversity and stem density both have a positive effect on LAI and a negative effect on percent DT. LAI decreased significantly with elevation by about 1.1 m2 m-2 per 1000 m altitude, reaching means of about 6.6 at 500 m and 5.2 at 2000 m a.s.l. DT increased significantly with elevation, revealing a tight negative correlation to LAI. Below canopy DT averaged 2.1% at 500 m and 4.0% at 2000 m (relative to incident radiation). Tree species diversity and stem density had a significant positive effect on LAI and a negative effect on DT. The observed LAI decrease with elevation is a consequence of both an altitudinal decrease in stem density and a decrease in tree species diversity combined with an effect of soil fertility (C/N ratio of the upper mineral soil and plant-available P of the organic layers). We also attempted to identify the key fa ctors controlling liana abundance. The abundance and diameters of all lianas (dbh ≥ 1 cm) and trees (dbh ≥ 10 cm) were recorded. Mean density of lianas and liana basal area did not differ significantly between the elevational levels; however, mean liana diameter decreased with increasing elevation. Liana density, basal area and abundance per tree individual were strongly correlated with host tree diameter. The most important determinants of liana abundance in the SBR were structural stand properties (tree basal area and mean tree dbh), but soil nitrogen availability (indicated by C/N ratio) was also found to have some influence. The notably high variation in liana abundance among different old-growth forest stands mainly reflects differences in stand structure between the studied mature forests, whereas the effect of elevation seems to be of less importance.