Environmental changes in biodiversity hotspot ecosystems of
South Ecuador: RESPonse and feedback effECTs (FOR2730)

Abstract:

The spatial heterogeneity of tropical forest epiphytes has rarely been quantified in terms of biomass. In particular, the effect of topographic variation on epiphyte biomass is poorly known, although forests on ridges and ravines can differ drastically in stature and exposure. In an Ecuadorian lower montane forest we quantified epiphytic biomass along two gradients: (1) the twig-branch-trunk trajectory, and (2) the ridge-ravine gradient. Twenty-one trees were sampled in each of three forest types (ridge, slope, ravine positions). Their epiphytic biomass was extrapolated to stand level based on basal area?epiphyte load relationships, with tree basal areas taken from six plots of 400 m 2 each per forest type. Our results document the successional addition and partial replacement of lichens by bryophytes, angiosperms and finally dead organic matter along the twig-branch-trunk trajectory. Despite having the highest tree basal area, total epiphytic biomass (mean ± SD) of ravine forest was significantly lower (2.6 ± 0.7 Mg half 1) than in mid-slope forest (6.3 ± 1.1 Mg half 1) and ridge forest (4.4 ± 1.6 Mg half 1), whereas maximum bryophyte water storage capacity was significantly higher. We attribute this pattern to differences in forest dynamics, stand structure and microclimate. Although our study could not differentiate between direct effects of slope position (nutrient availability, mesoclimate) and indirect effects (stand structure and dynamics), it provides evidence that fine-scale topography needs to be taken into account when extrapolating epiphytic biomass and related matter fluxes from stand-level data to the regional scale.

Abstract:

Microbial oxidation in aerobic soils is the primary biotic sink for atmospheric methane (CH4), a powerful greenhouse gas. Although tropical forest soils are estimated to globally account for about 28% of annual soil CH4 consumption (6.2 Tg CH4 year−1), limited data are available on CH4 exchange from tropical montane forests. We present the results of an extensive study on CH4 exchange from tropical montane forest soils along an elevation gradient (1,000, 2,000, 3,000 m) at different topographic positions (lower slope, mid-slope, ridge position) in southern Ecuador. All soils were net atmospheric CH4 sinks, with decreasing annual uptake rates from 5.9 kg CH4?C ha−1 year−1 at 1,000 m to 0.6 kg CH4?C ha−1 year−1 at 3,000 m. Topography had no effect on soil atmospheric CH4 uptake. We detected some unexpected factors controlling net methane fluxes: positive correlations between CH4 uptake rates, mineral nitrogen content of the mineral soil and with CO2 emissions indicated that the largest CH4 uptake corresponded with favorable conditions for microbial activity. Furthermore, we found indications that CH4 uptake was N limited instead of inhibited by NH4 +. Finally, we showed that in contrast to temperate regions, substantial high affinity methane oxidation occurred in the thick organic layers which can influence the CH4 budget of these tropical montane forest soils. Inclusion of elevation as a co-variable will improve regional estimates of methane exchange in these tropical montane forests.

Abstract:

Tropical forests are important sources of the greenhouse gas nitrous oxide (N2O) and of nitric oxide (NO), a precursor of ozone. In tropical montane forests nitrogen limitation is common which affects both soil N2O and NO fluxes and forest productivity. Here we present evidence that forest productivity and N-oxide (N2O + NO) fluxes are linked through N availability along elevation and topographic gradients in tropical montane forests. We measured N-oxide fluxes, several indices of N availability, and forest productivity along an elevation gradient from 1000 m to 3000 m and along topographic gradients. Organic layer thickness of the soils increased and N availability decreased with increasing elevation and along the topographic gradient from the lower slope position to the ridges. Annual N2O fluxes ranged from -0.53 µg(N)m-2h-1 to 14.54 µg(N)m-2h-1 while NO fluxes ranged from -0.02 µg(N)m-2h-1 to 1.13 µg(N)m-2h-1. Both N-oxide fluxes and forest productivity increased with increasing N availability and showed close positive correlations with indices of N availability (C/N ratio and  15N signature of litterfall). We interpret the close correlations of N-oxide fluxes with total litterfall and tree basal area increment as evidence that N availability links N-oxide fluxes and forest productivity. This opens the possibility to include forest productivity as co-variable in predictions of N-oxide fluxes in nitrogen limited tropical montane forests. Especially increment of tree basal area was a promising proxy to predict soil N-oxide fluxes in these N limited ecosystems, possibly because it better reflects long-term forest productivity than total litterfall.

Abstract:

High geomorphic process rates often coincide with the sensitivity of (tropical) mountain eco-systems towards both man-made and naturally controlled environmental changes. Especially the pastured tertiary sediments of the semiarid western Andean slope in Southern Ecuador are characterized by a higher occurrence of gravitational mass movements, reaching the point where the landscape is about to change into "badlands". Based on the analysis of the landslide inventory of a representative study area near Masanamaca the present work dealt with the reasons for these increased geomorphic process rates. The study area stretches over 10 km² and is built up by tertiary sediments in the west and by metamorphic rock in the east. A se-cond aim was the comparison with the study area Reserva Biológica San Francisco (RBSF) on the perhumid, metamorphic slope of the Andes because of major differences regarding precipitation, land use and types of movements. Both study areas are located within a distance of about 35 km. First of all 900 landslides in the study area Masanamaca were mapped and classified with respect to type of movement, material, activity and causing factors. Five different major types of movement were found (438 translational and 25 rotational slides, 63 flow ~, 162 complex ~, 156 composite ~ and 56 fall movements). The slope angle and discontinuities in the bed-rock formed the main preparatory factors which predisposed the slope to failure. In this re-spect the lithology exerted a special influence as well as the biggest landslide density was observed in the tertiary sediments (conglomerate and sand-/siltstone). Precipitation represent-ed the most common triggering factor, a fact which was closely related to the predominant occuerrenc of shallow movements (slides, flows and some complex movements) in the study area which, as widely acknowledged, are mostly rainfall-induced. Especially the initation of large-scale and deep-seated movements such as composite landslides must have required an array of triggering factors. They might also represent an advanced stage of ongoing surface lowering activity which could have had its origin in a small slide scar. The calculation of the denudation rate caused by shallow sliding demanded in the first place the estimation of the displaced landslide volume for both study areas. This was achieved by using area-volume scaling. Secondly, for the study area Masanamaca it was necessary to es-timate the time needed for the regeneration of the observed shallow landslide population. For this purpose thresholds proposed in the literature were used to assess roughly the frequency of landslide triggering rain events in the study area. The literature review in combination with the analysis of the precipitation in the study area resulted in an estimated time span between 10 to 100 years, which corresponded to an annual denudation rate between 0.5 to 5 mm. By con-trast, the analysis of time-lapse airborne imagery for the study area RBSF for six different years made estimations of recurrence intervals unnecessary. For a 38-year period (1962-2000) a denudation rate of 2 mm/y caused by naturally induced landslides and one of 12 mm/y caused by man-made triggered landslides were calculated. All landslides of the study area RBSF were classified as shallow movements. In the study area Masanamaca however much more types of movements were identified. Nevertheless a great number of them were disregarded in the calculation of the denudation rate as neither volume nor recurrence time could be estimated for them. Since these unconsidered movements were both large-scale and deep-seated, it can be assumed that their consideration would have led to a denudation rate potentially higher than the naturally constituted one of the perhumid Andean slope. Presented results can be the basis of further investigations which might focus on the analysis of susceptibility towards shallow landsliding and associated hazards. For this purpose a more suitable denudation rate for the study area Masanamaca would be preferable. This could be achieved by implementing empircal or physical modeling. ²denudation rate = displaced volume/(3-D area of the study area * time)

Abstract:

Soil-vegetation-atmosphere transfer (SVAT) is to be predicted for 2050/2100 for a study area in the southern Ecuadorian Andes. SVAT models require information on land use/ land cover (LUC) as lower boundary conditions. Since the study area suffers from high deforestation rates, LUC cannot be assumed as staying constant with time. A spatially explicit land use/land cover change (LUCC) model is therefore needed for future SVAT prediction. The numbers of approaches of LUCC modelling are numerous. Difficulties are due to complex interactions of social and biophysical drivers of change. In this study a model of LUCC was built using information of past changes derived by remotely sensed data. Special focus was on forest development patterns. A training period of 14 years between 1987 and 2001 was chosen. Two LUC classifications were accomplished to Landsat data of the start and end date of this period. A change detection of the training period provided the basis for predictive LUCC modelling. Potential drivers for LUCC were applied to the model as GIS layers. The modelling procedure consisted of a combination of Markov chain analysis (MCA) for quantitative modelling and multi-layer perceptron (MLP) for revealing potential locations of change. A multi-objective land allocation (MOLA) served as final integration step. 14 LUC transitions were considered in the modelling procedure. Unconsidered LUC classes were assumed to stay constant in the future. The model results were maps of LUC for 2006, 2010 and afterwards for every 10 years up to 2100. An internal validation was performed with the training data. The results of the prediction were validated by comparing the model output of 2006 to an ASTER LUC classification of the same time. The validation methodology comprised crisp and fuzzy map comparison using Kappa statistics. The study area featured a deforestation of 13.61% in the training period. The model was able to explain deforestation in the training period 51% better than just by chance. The location of predicted deforestation reached a better than chance agreement of 30%. Predicted quantities of deforestation were 59% conforming with the reference. The validation of the prediction indicated the difficulty of modelling human impact on the ecosystem. Prospects and limitations of the model were identified with suggestions for future research tasks. The results of this study are assumed to present a good groundwork for future SVAT models.

Abstract:

Caterpillar communities were studied on 16 shrub species from the family Asteraceae and the genus Piper (Piperaceae) in the Andean montane rainforest zone of southern Ecuador. The applied sampling methodology was evaluated and found well suited for sampling of caterpillars, especially with focus on resampling. During standardized sampling, a total of 18890 specimens were collected in 11 field surveys between August 2007 and June 2009. Overall, samples were dominated by gregarious early instars and egg clutches of Altinote dicaeus albofasciata (Nymphalidae) on Erato polymnioides (Asteraceae). Those and all other such groups and egg clutches were downweighted for analyses to their cubic root to balance against their overrepresentation in the data set. Trophic associations of caterpillars were confirmed by feeding trials. A substantial fraction of more than 22% of the encountered caterpillars (and up to 80% on individual shrub species) was found not to be trophically linked to the living biomass of the shrubs themselves (“non-herbivores”), but rather feed on dead leaves and epiphylls. Abundance of nonherbivores differed strongly between the two studied plant families, but was very similar on different shrub species within these families. Abundance of herbivorous caterpillars, to the contrary, differed hardly between plant families, but varied strongly between individual plant species. Herbivores were almost entirely comprised of ectophagous folivores, while florivores (2.3%) and semi-endophagous folivores (leaf rollers and tiers; 6.0%) were unexpectedly rare. Absolute species richness of herbivorous caterpillars was high, with a total of 191 Lepidoptera species on the studied 16 shrub species, but varied up to 40fold between individual plant species. Rarefied species numbers were more similar among plant species, but still showed a 15fold difference between the most species rich and most species poor community. Communities on Piper species were characterized by low effective species numbers (measured as exponential Shannon entropy) and high dominance of one or two species of the Geometridae genus Eois. E. polymnioides featured a similar structure dominated by A. dicaeus albofasciata. Communities on the other two Asteraceae (Ageratina dendroides and Baccharis latifolia), to the contrary, were found to have high effective species numbers and low dominance. In conclusion, while Piper species and E. polymnioides feature caterpillar communities defined by small, predictable sets of core herbivores, these defining sets do not exist for the other two of the studied Asteraceae species. Communities on Piper were also more specialized, with 62.8% of the caterpillar species considered monophagous at plant species level, than on Asteraceae with only 11.6% monophagous species. The observed diversity patterns point mainly to shaping by (chemical) plant-herbivore-defense, while geographic range and local abundance of host plants appear to have only limited and varying influence on the associated caterpillar communities.

Abstract:

The genus Eois comprises an important part of megadiverse assemblages of geometrid moths in mountain rainforests of southern Ecuador. In this study we report: (i) on the construction of a DNA barcode library of Eois for identification purposes; and (ii) the exploration of species diversity through species delimitation by pair-wise distance thresholds. COI barcode sequences were generated from 408 individuals (at least 105 species) collected on a narrow geographic scale (∼40 km2) in the Reserva Biol´ogica San Francisco. Analyses of barcode sequence divergence showed that species delimitations based solely on external morphology result in broad overlap of intra- and interspecific distances. Species delimitation at a 2% pair-wise distance threshold reveals a clear barcoding gap. Fifty-two previously unrecognized species were identified, 31 of which could only be distinguished by an integrative taxonomy approach. Twelve additional putative species could only be recognized by threshold-based delimitation. Most splits resulted in two or three newly perceived cryptic taxa. The present study increased the number of Eois species recorded from that small area of Andean mountain forest from 102 to 154 (morphology- plus integrative taxonomy-based) or even 166 (sequence-based), leaving the species accumulation curve still far from reaching an asymptote. Notably, in no case did two or more previously distinguished morphospecies have to be lumped. This barcode inventory can be used to match larvae to known adult samples without rearing, and will therefore be of vital help to extend the currently limited knowledge about food plant relationships and host specialization.

Abstract:

We analyzed the stocks of Ca, Mg, P, and S in tree rings of Cedrela montana Moritz ex Turcz in order to evaluate the dynamics of element input to a lower montane rain forest in South Ecuador. Samples were taken by a Pressler drill from the living tree (n=5). Afterwards tree rings were separated and digested with concentrated HNO3 under pressure (Heinrichs et al.,1986). Concentrations of Ca and Mg were determined by atomic absorption spectroscopy (AAS), concentrations of P and S by inductively coupled plasma optical emission spectrometry (ICP-OES), respectively. Mean element concentrations in tree rings between 1957 and 2005 were 2.2 g kg-1 for Ca, 1.1 g kg-1 for P and S and 0.4 g kg-1 for Mg. Radial distribution of element concentrations was heterogeneous with exception of the tree rings close to the phloem, where concentrations were generally higher. The pattern of annual radial increment showed significant growth peaks related to strong La Niña events. Also concentrations of P and S peaked during the La Niña events of 1989, 1996, and 2000. For Ca and Mg a similar effect was observed with a time shift of one year. Highest Ca peaks during the monitored period were associated with the 1974 La Niña event, which was also the strongest event on the record. We conclude that tree rings are a suitable tool for recording element inputs of Ca, Mg, P, and S at lower montane rain forests of South Ecuador. Furthermore we found element inputs and growth signals linked to the ENSO cycle.

Abstract:

Catchment-scale runoff generation involves a complex interaction of physical and chemical processes operating over a wide distribution of spatial and temporal scales. Understanding runoff generation is challenged by this inherent complexity ? the more uncertain step of predicting the hydrologic response of catchments is that much more challenging. Many different hypotheses have been implemented in hydrological models to capture runoff generation processes and provide hydrologic predictions. These concepts have been developed based on extended field observations. Here we propose inferring water flux understanding and catchment exploring through the application of a variety of available hydrological models as a mechanism to build upon and extend models that have been developed to capture particular hydrological processes. We view this ensemble modeling strategy as particularly appropriate in ungauged catchments. The study is carried out in a tropical montane rainforest catchment in Southern Ecuador. The catchment is 75 km2 and is covered by forest in the south, while the northern slopes have been partly deforested for grazing. Annual rainfall is highly variable, reaching up to 5700 mm per year in the upper parts of the catchment. To explore the dominating runoff processes, an ensemble of 6 hydrological models with different structures applied over different levels of both spatial and temporal detail was developed. The ensemble includes spatially lumped (HBV-light), semi-distributed (HEC-HMS, CHIMP, SWAT, LASCAM) and a fully distributed model (HBV-N-D). The hydro-statistical toolkit WETSPRO was used to characterize simulated and observed hydrographs. Estimated baseflow indices, flow minima and maxima, flow duration curves and cumulative errors were generated and compared among the ensemble of models. This process facilitated the exploration of processes controlling runoff generation, enabled an evaluation of the applicability of the screened models to tropical montane rainforests, and provided the capacity to evaluate and explain where different models failed.

Abstract:

The understanding of the hydrological functioning of a tropical ecosystem allows to evaluate the consequences of a changed land use. On the east-facing slope of the Andes in Ecuador between 1900 and 2150 m a.s.l. I studied the above- and belowground driving variables of the water budget. The objectives were to determine the elements of the water budget experimentally and with an optimized model and to elucidate the hydraulically relevant processes along five 20 m-long transects in three 8-13 ha-large microcatchments under montane rain forest. Water fluxes were weekly measured between April 1998 and April 2001, the mean annual evapotranspiration was calculated with a budget approach. For the canopy, I determined the water storage capacity and assessed the influence of the vegetation density and epiphyte coverage on interception and canopy evaporation of the precipitation. The vegetation density was approximated by the leaf area index [LAI] and the fraction of the precipitation falling through gaps of the canopy. In addition, I modeled throughfall and stemflow by adapting the analytical canopy model of Gash. In the soil, I determined the saturated hydraulic conductivity and the water contents. To calculate the reaction of surface flow on precipitation events and to identify the soil processes controlling the surface flow, I compared a ?black box? (NAMOD) with a ?semi-empirical? (TOPMODEL) model approach. The mean water storage capacity of the dry leaves is 1.91±1.94 mm and that of the trunks 0.041±0.02 mm. The LAI ranges from 5.2 to 9.3. Epiphytes - mainly mosses - cover up to 80% of the trunk and branch surfaces. The fraction of precipitation falling through gaps of the canopy correlates significantly with the interception loss (r = -0.77, n = 40). During dry periods the mosses and lichens tend to decrease the water storage capacity of the trunks. The saturated hydraulic conductivity decreases drastically at the border between organic layer and mineral soil where it was highly variable. The high stone content of, on average, 40Vol.% has a substantial effect of the flow regime. Based on the results of the precipitation-surface flow modeling the hydraulically effective soil layer can be separated into two water reservoirs (organic layer and mineral soil). The water flow in the soil occurres rapidly in lateral direction in the organic layer and/or much slower, mainly vertically in the deeper mineral soil layers. Water flow in the latter soil layers during strong rain events is little likely, because the organic layer is rapidly saturated favoring interflow. The weekly surface flow measurements result in an overestimation of the annual water output from the catchments because high flow events were overrepresented. The mean annual surface flow in the three catchments, calculated on the basis of daily modeled flow rates, ranges between 27 and 46% of the precipitation. The annual evaporation, calculated on the basis of daily surface flow rates, ranges between 1281 and 1889 mm, i.e. 54-73% of the precipitation. My results demonstrate that rapid lateral water flow in the organic layer of the soil dominates the flow regime. The destruction of the forest and removal of the organic layer would result in soil erosion and increased intensity of high-flow events.

Abstract:

Tropical mountain regions are characterized by strong spatial climate gradients which together with the limited amount of data and knowledge of the underlying processes hinder the management of the water resources. Especially for regional-scale prediction it is important to identify the dominant factors controlling the rainfall?runoff response and link those to known spatial patterns of climate, soils, and vegetation. This study analyzes the rainfall?runoff relation of 13 intensively monitored micro-catchments in the Andes of southern Ecuador. The results of this study show that streamflow in the southern cordillera of the Ecuadorian Andes, above 2500 m a.s.l., primarily consists of subsurface flow. The yearly amount of streamflow is controlled by the annual rainfall depth, whereas the temporal distribution is mainly governed by the lateral saturated hydraulic conductivity, the soil water retention and the antecedent soil moisture content. Anthropogenic effects were found insignificant, with the exception in one of the studied micro-catchment. Effect of land use changes in most of the micro-catchments did not reflect in the shape of the flow duration curve because either the spatial extent of human impact was small and/or the overall basin slope was less than 20%.

Abstract:

The near surface humidity in a megadiverse mountain ecosystem in southern Ecuador is examined on the basis of Relative Humidity (RH) measurements inside the natural mountain forest and at open sites along an altitudinal gradient from 1700 to 3200 m. The main methodological aim of the current study is to generate a humidity regionalization tool to provide spatial datasets on average monthly mean, minimum and maximum RH, Specific Humidity (q) and Specific Saturation Deficit (DS) by using observation data of RH. The maps based on data of the period 1999–2009 are needed by ecological projects working on various plots where no climate station data are available. The humidity maps are generated by combining a straightforward detrending technique with a Digital Elevation Model and a satellite-based land cover classification which also provides the relative forest cover per pixel. The topical aim of the study is to investigate the humidity distribution and structure of both manifestations of our ecosystem (pastures and natural vegetation) with special considerations to the ecosystem regulation service by converting natural forest into pasture. The results reveal a clear differentiation over the year, partly triggered by the change of synoptic weather situation but also by land cover effects. Humidity amplitudes are particularly low during the main rainy season when cloudiness and rainfall are high, but markedly pronounced in the relative dry season when daily irradiance and outgoing nocturnal radiation causes distinct differences between the land cover units. Particularly the upper pasture areas gained by slash and burn of the natural forest exhibit the lowest humidity values while the humidity inside the mountain forest is significantly higher due to the regulating effects of the dense vegetation. Thus, clearing the forest clearly reduces the regulation function (regulating ecosystem services) of the ecosystem which might become problematic for reforestation under future global warming.

Abstract:

Stem diameter increments of the broadleaved deciduous tree species Tabebuia chrysantha were measured with high-resolution dendrometers in a tropical lower montane forest and in a dry forest in southern Ecuador, the latter showing a distinct dry season. Those analyses were complemented by wood anatomical studies on regularly collected microcores to determine the season of active cambial growth and the time of formation of annual growth boundaries. The length of the cambial active period varied between 3 and 7 months at the tropical lower montane forest and 2 and 4 months in the dry forest, respectively. During dry days, amplitudes of daily stem diameter variations correlated with vapour pressure deficit. During October and November, inter-annual climate variations may lead to dry and sunny conditions in the tropical lower montane forest, causing water deficit and stem diameter shrinkage in T. chrysantha. The results of the climate– growth analysis show a positive relationship between tree growth and rainfall as well as vapour pressure deficit in certain periods of the year, indicating that rainfall plays a major role for tree growth.

Abstract:

This diploma thesis deals with the purpose to create a dataset of soil texture (sand and clay content) as input parameter for a SWAT model. For this a senstitvity study on this parameter in the model was carried out and a statistical approach based on a digital terrain analysis to regionalize soil profile data was driven. The results show that the sensitivity of this model (CLM) limits to extreme content ratios of sand and clay. The resulting maps of the regionalization is discussed critical due to various factors: i) the number of used soil profile samples is to small ii) the accuracy of the DEM is to uncertain iii) the surface of the model region is highly textured (step slopes) and it should be considered wether the pedological processes are sorting soil substrate like in moderate regions, for which the concept of deriving soil properties in dependency of the terrain was proven.

Abstract:

Soil-landscapes are diverse and complex due to the interaction of pedogenetic, geomorphological and hydrological processes. The resulting soil profile reflects the balance of these processes in its properties. Early conceptual models have by now resulted into quantitative soil-landscape models including soil variation and its unpredictability as a key soil attribute. Soils in the Andean mountain rainforest area of southern Ecuador are influenced by hillslope processes and landslides in particular. The lack of knowledge on the distribution of soils and especially physical soil properties to understand slope failure, resulted in the study of this particular soil-landscape by means of statistical models relating soil to terrain attributes, i.e. predictive soil mapping. A 24 terrain classes comprising sampling design for soil investigation in mountainous areas was developed to obtain a representative dataset for statistical modelling. The soils were investigated by 56 profiles and 315 auger points. The Reference Soil Groups (RSGs) Histosol, Stagnosol, Umbrisol, Cambisol, Leptosol and Regosol were identified according to the World Reference Base for Soil Resources (WRB). Classification tree models and a probability scheme based on WRB hierarchy were applied to include RSG prediction uncertainty in a digital soil map. Histosol probability depended on hydrological parameters; highest Stagnosol probability was found on slopes < 40° and above 2146 m a.s.l. Poor model performance, probably due to the prediction of complex categories (RSGs) and WRB inconsequence (absolute and relative value criteria), led to the proposal of ?incomplete soil classification? by relating the thickness of the WRB?s diagnostic horizons as percentage to the upper 100 soil centimetres, including the organic layer. Typical diagnostic horizons histic, humic, umbric, stagnic and cambic were regionalised in their thickness and occurrence probability by classification and regression trees (CART). Prediction uncertainty was addressed with hundredfold model runs based on different random Jackknife partitions of the dataset. Whether the first mineral soil horizon displays stagnic properties or not, likely depends on physical soil properties in addition to terrain parameters. Incomplete soil classification resulted in histic and stagnic soil parts dominating the first 100 cm of the soil volume for most of the research area. While soil profiles and auger points were described in their horizon composition, thickness, Munsell colour and soil texture by finger method (FAO, 2006), soil cohesion, bulk density and texture by pipette and laser were analysed in soil profiles only. Texture results by pipette compared to laser method, showed the expected shift to higher silt and lower clay contents. Linear regression equations were adapted. Pedotransfer functions to predict physical soil properties from the bigger auger dataset analysed by field texture method only, could not be developed, because field texture analysis did not provide satisfying results. It was therefore not possible to correct its results with the more precise laboratory data. Comparing CART and Random Forest (RF) in their model performance to predict topsoil texture and bulk density as well as mineral soil thickness by hundredfold model runs with random Jackknife partitions, RF predictions resulted more powerful. Altitude a.s.l. was the most important predictor for all three soil parameters. Increasing sand/ clay ratios with increasing altitude, on steep slopes and with overland flow distance to the channel network are caused by shallow subsurface flow removing clay particles downslope. Deeper soil layers are not influenced by the same process and therefore showed different texture properties. Terrain parameters could only explain the spatial distribution of topsoil properties to a limited extent, subsoil properties could not be predicted at all. Other parameters that likely influence soil properties within the investigation area are parent material and landslides. Strong evidence was found that topsoil horizons did not form from the bedrock underlying the soil profile. Parent material changes within short distance and often within one soil profile. Landslides have a strong influence on soil-landscape formation in shifting soil and rock material. Soil mechanical and hydrological properties in addition to terrain steepness were hypothesized to be the major factors in causing soil slides. Thus, the factor of safety (FS) was calculated as the soil shear ratio that is necessary to maintain the critical state equilibrium on a potential sliding surface. The depth of the failure plane was assumed at the lower boundary of the stagnic soil layer or complete soil depth, depending on soils being stagnic or non-stagnic. The FS was determined in dependence of soil wetness referring to 0.001, 0.01, 0.1 and 3 mm/h net rainfall rate. Sites with a FS &#8805; 1 at 3 mm/h (complete saturation) were classified as unconditionally stable, sites with a FS < 1 at 0.001 mm/h as unconditionally unstable. The latter coincided quite well with landslide scars from a recent aerial photograph.

Abstract:

Over the past decades, the tropical mountain rainforest of southern Ecuador has been threatened by conversion to cattle pastures. Frequently, these pastures are invaded by bracken fern and abandoned when bracken becomes dominant. Changes in land-use (forest–pasture–abandoned pasture) can affect soil microorganisms and their physiological responses with respect to soil carbon and nutrient cycling. In situ investigations on litter decomposition and soil respiration as well as biogeochemical characterization of the soil were carried out to identify the driving factors behind. The conversion of forest to pasture induced a pronounced increase in CO2–C effluxes to 12.2 Mg ha-1 a-1 which did not decrease after abandonment. Soil microbial activity and biomass showed a different pattern with lowest values at forest and abandoned pasture sites. With 3445 mg kg-1 (0–5 cm) microbial biomass carbon (MBC by CFE-method), the active pasture had a more than three times higher value than forest and abandoned pasture, which was among the highest in tropical pasture soils. A shift in the microbial community structure (phospholipid fatty acid, PLFA) was also induced by the establishment of pasture land; the relative abundance of fungi and Gram-negative bacteria increased. PLFA fingerprints of the forest organic layer were more similar to pasture than to forest mineral soil. Chemical properties (pH value, exchangeable cations) were the main factors influencing the respective microbial structure. Bracken-invasion resulted in a decrease in the quantity and quality of aboveand belowground biomass. The lower organic substance and nutrient availability induced a significant decline in microbial biomass and activity. After pasture abandonment, these differences in soil microbial function were not accompanied by pronounced shifts in the community structure and in soil pH as was shown for the conversion to pasture. A disconnection between microbial structure and function was identified. Similar soil CO2–C effluxes between active and abandoned pasture sites might be explained by an underestimation of the effluxes from the active pasture site. All measurements were carried out between grass tussocks where fine-root density was about 2.6 times lower than below tussocks. Thus, lower proportions of root respiration were expected than below tussocks. Overall, soil microorganisms responded differently to changes in land-use from forest to pasture and from pasture to abandoned pasture resulting in pronounced changes of carbon and nutrient cycling and hence of ecosystem functioning.