Search

Assessing large scale plant productivity of coastal marshes is essential to understand the resilience of these systems to climate change. Two machine learning approaches, random forest (RF) and support vector machine (SVM) regression were tested to estimate biomass of a common saltmarshes species, salt couch grass (Sporobolus virginicus). Reflectance and vegetation indices derived from 8 bands of Worldview-2 multispectral data were used for four experiments to develop the biomass model. These four experiments were, Experiment-1: 8 bands of Worldview-2 image, Experiment-2: Possible combination of all bands of Worldview-2 for Normalized Difference Vegetation Index (NDVI) type vegetation indices, Experiment-3: Combination of bands and vegetation indices, Experiment-4: Selected variables derived from experiment-3 using variable selection methods. The main objectives of this study are (i) to recommend an affordable low cost data source to predict biomass of a common saltmarshes species, (ii) to suggest a variable selection method suitable for multispectral data, (iii) to assess the performance of RF and SVM for the biomass prediction model. Cross-validation of parameter optimizations for SVM showed that optimized parameter of ɛ-SVR failed to provide a reliable prediction. Hence, ν-SVR was used for the SVM model. Among the different variable selection methods, recursive feature elimination (RFE) selected a minimum number of variables (only 4) with an RMSE of 0.211 (kg/m2). Experiment-4 (only selected bands) provided the best results for both of the machine learning regression methods, RF (R2= 0.72, RMSE= 0.166 kg/m2) and SVR (R2= 0.66, RMSE = 0.200 kg/m2) to predict biomass. When a 10-fold cross validation of the RF model was compared with a 10-fold cross validation of SVR, a significant difference (p = <0.0001) was observed for RMSE. One to one comparisons of actual to predicted biomass showed that RF underestimates the high biomass values, whereas SVR overestimates the values; this suggests a need for further investigation and refinement. [ABSTRACT FROM AUTHOR]

Controls on the characteristics of floodplain wetlands in drylands are diverse and may include extrinsic factors such as tectonic activity, lithology and climate, and intrinsic thresholds of channel change. Correct analysis of the interplay between these controls is important for assessing possible channel-floodplain responses to changing environmental conditions. Using analysis of aerial imagery, geological maps and field data, this paper investigates floodplain wetland characteristics in the Tshwane and Pienaars catchments, northern South Africa, and combines the findings with previous research to develop a new conceptual model highlighting the influence of variations in aridity on flow, sediment transport, and channel-floodplain morphology. The Tshwane-Pienaars floodplain wetlands have formed in response to a complex interplay between climatic, lithological, and intrinsic controls. In this semi-arid setting, net aggradation (alluvium >7 m thick) in the wetlands is promoted by marked downstream declines in discharge and stream power that are related to transmission losses and declining downstream gradients. Consideration of the Tshwane-Pienaars wetlands in their broader catchment and regional context highlights the key influence of climate, and demonstrates how floodplain wetland characteristics vary along a subhumid to semi-arid climatic gradient. Increasing aridity tends to be associated with a reduction in the ability of rivers to maintain through-going channels and an increase in the propensity for channel breakdown and floodout formation. Understanding the interplay between climate, hydrology and geomorphology may help to anticipate and manage pathways of floodplain wetland development under future drier, more variable climates, both in South African and other drylands. Copyright © 2016 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

1. Many plants disperse their seeds in waterbodies via hydrochoric transport. Despite a growing body of research into hydrochory, little is known about the fundamental seed traits that determine floatation ability or hydrochoric transport behaviour more generally. Seeds are transported in fluvial systems in one of three phases: surface transport, within the flow or incorporated in bedload. Seeds are often categorised as buoyant or non-buoyant based on density, with little consideration of the morphological traits that determine how long seeds spend in each transport phase. 2. We investigated the seed traits that best predict time spent floating under laboratory conditions, using sixty species of riparian plants from south-eastern Australia. We measured and calculated key physical attributes (length, width, volume, surface area, mass, density, sphericity, roundness and shape category) and categorised seeds according to their primary dispersal modes (water/wind, vertebrate, adhesion, ants and unassisted). 3. We used Cox's proportional hazards modelling to reveal that seed density and volume : surface area ratio were the physical traits that best predicted time spent floating. Volume : surface area ratio represents both seed size and shape, as it increases with size and sphericity. Coir results showed that denser, larger and/or more spherical seeds (i.e. higher volume : surface area ratio) were poor floaters. They are therefore more likely to be transported in subsurface flow or as bedload. 4. We combined our data with geomorphic models of sediment transport in flow to develop a biophysical process model of how hydrological forces and seed physical traits determine hydrochorous seed transport behaviour in rivers. The model describes how surface tension, buoyancy and flow velocity act on seed density and volume : surface area ratio to characterise time spent floating. We extrapolate from our data to conceptualise how these traits predict transitions between surface, subsurface and bedload transport. Hydrochoric seed transport behaviour and deposition are both threshold-driven and multidirectional. 5. Coir process model is likely to be applicable across a range of different seed types in a range of rivers. It substantially increases our understanding of hydrochoric processes in rivers, lakes and lotic wetlands and will help illuminate the links between seed traits, hydrochoric transport and patterns of riparian vegetation and species composition. [ABSTRACT FROM AUTHOR]