Interpreting reach-scale classifications and the role of spatial-morphological variables in river channel mapping using machine learning algorithms

<p>Over the years, there has been tremendous growth in the literature as regards river channel classifications, however, very few studies have been able to engage the use of remote sensing products in channel classification at the reach-scale level especially by combining reflections from satellite sensors with channel morphological variables. This study aims to identify discriminating spatio-morphological variables using machine learning algorithms and classify site-specific channel types at the reach scale. Each reach was broadly classified based on valley settings (confined, partly confined and unconfined) and channel types (alluvial or bedrock). However, variations and site observations were recorded for site-specific classification purposes. For each reach, Global Positioning System devices were used to geo-locate their endpoints. Standard field instruments were used for cross-sectional measurements and established hydraulic equations for the derived variables. A total of 249 points across 83 reaches were sampled during the fieldwork. Landsat 8 and Sentinel-1 bands were retrieved for days the fieldwork was carried out/for days close to those dates using Google Earth Engine (GEE) platform. Hierarchical cluster analysis, HCA, using Ward’s linkages was used to provide a classification for the channel types. For the identification of important variables in predicting channel unit types, the random-forest - recursive feature elimination (RF-RFE) algorithm was used using the rfe() function. To identify the best machine learning algorithm, random-forest (rf), support vector machines (svm), multivariate adaptive regression spline (mars) extreme gradient boosting (xgb) and adaptive boosting (adaboost) were used on the training and test data to identify the best performing algorithm. The rfe() feature selection identified five (5) variables that can significantly help in channel unit type identification. The top five variables are dimensionless stream power, slope, width, wetted perimeter and Band 4. Using ROC curve, sensitivity, and specificity, the mars model has the highest ROC curve. Hence, it appears to be the best performing out of the five. However, if the argument is to be based on positive prediction, then any of the models except adaboost will be preferred given their high sensitivity. The HCA using illustrated the clustering structure of the studied reaches by producing five distinct channel classification types distinguished based on width-depth ratio values (high and low). The five distinct channel types are listed as  M1e, M5e, B1, E5b, and E. These codings are based partly on Rosgen’s classification while, the capital letters (M, B and E) represent mixed channels, bedrock with moderate width-depth ratio and alluvial channels with low width-depth ratio respectively. Numbers 1 and 5 represent bedrocks and sandy beds based on slope variation respectively. The identified channel unit types are a result of the underlying lithology, process-form dynamics and confinement. As streams are expected to respond differently to shocks and recover from damages, it becomes essential to understand these differences in classification which will go a long way in establishing watershed and streamside management guidelines.</p>

High-Throughput Synthesis of Liposome Using an Injection-Molded Plastic Micro-Fluidic Device

For mass production of liposomes, we designed a plastic micro-channel device on the basis of 5 μm of micro-nozzle array forming T-junction with 100 μm depth of micro-channel. A micro-channel unit for synthesizing liposomes consisted of two micro-nozzle arrays for mixing two solutions as well as delivery and recovery channels for supplying solutions and collecting liposome suspension. The number of micro-nozzles was approximately 2400 for a micro-channel unit, and seven units were applied independently on a micro-channel plate. The plastic micro-channel plate was injection-molded for mass production using a micro-channel stamper previously fabricated by UV lithography and nickel electroforming process. A plastic cover plate with seven pairs of inlet and outlet ports was machined by mechanical milling and drilling and was assembled with a micro-channel plate using a holder to form a liposome synthesizing device. Flow and mixing of solutions in the micro-channels were tested using colored water to check the micro-fluidic characteristics of the device. Finally, a L-α-phosphatidylcholine (SOY PC) liposome was synthesized using EtOH solution of SOY PC (95%) and saline (0.85% NaOH solution) to find that the liposomes were around 230 and 260 nm in diameter, depending on the flow rate of the lipid solution.

Time Series Changes in Fish Assemblages and Habitat Structures Caused by Partial Check Dam Removal

Despite a steep increase in dam removal projects, there are few studies on the biophysical responses to dam removal. In this study, we evaluated the short-term (1.5 years) and long-term (5.5 years) effects of partial check dam removal on fish assemblages and their habitats. First, the habitat preferences of fish were examined at a channel unit scale: pools, rapids, and side channels satisfied the seasonal habitat requirements of the fish assemblages. Partial check dam removal increased the abundance of these habitats and diversified the habitat structures owing to the sediment release from the dam; in contrast, the bedrock distinctly decreased 1.5 years after dam removal. However, 5.5 years after dam removal, the bedrock proportion increased, and the habitat structures were simplified again owing to the re-transportation of sediments. Similar temporal changes were also determined through a reach-scale analysis with a change in the bed materials. Anadromous Oncorhynchus masou began to spawn and recolonize in the upstream section of the dam after the dam removal, causing similar assemblage compositions between the downstream and upstream sections. The abundance of Salvelinus malma and Noemacheilus barbatulus toni in the upstream reaches decreased over time after the dam removal. The temporal changes in the biophysical responses suggest that long-term monitoring is indispensable for the reliable evaluation of dam removal effects.

Characterization of morphological units in a small, forested stream using close-range remotely piloted aircraft imagery

Forested, gravel-bed streams possess complex channel morphologies which are difficult to objectively characterize. The spatial scale necessary to adequately capture variability in these streams is often unclear, as channels are governed by irregularly spaced features and episodic processes. This issue is compounded by the high cost and time-consuming nature of field surveys in these complex fluvial environments. In larger streams, remotely piloted aircraft (RPA) have proven to be effective tools for characterizing channels at high resolutions over large spatial extents, but to date their use in small, forested streams with closed forest canopies has been limited. This paper seeks to demonstrate an effective method for classifying channel morphological units in small, forested streams and for providing information on the spatial scale necessary to capture the dominant spatial morphological variability of these channels. This goal was achieved using easily extractable data from close-range RPA imagery collected under the forest canopy (flying height of 5–15 m above ground level; ma.g.l.) in a small (width of 10–15 m) stream along its 3 km of salmon-bearing channel. First, the accuracy and coverage of RPA for extracting channel data were investigated through a subcanopy survey. From these survey data, relevant cross-sectional variables (hydraulic radius, sediment texture, and channel slope) were extracted from high-resolution point clouds and digital elevation models (DEMs) of the channel and used to characterize channel unit morphology using a principal component analysis-clustering (PCA-clustering) technique. Finally, the length scale required to capture dominant morphological variability was investigated from an analysis of morphological diversity along the channel. The results demonstrate that subcanopy RPA surveys provide a viable alternative to traditional ground-based survey approaches for mapping morphological units, with 87 % coverage of the main channel stream bed achieved. The PCA-clustering analysis provided a comparatively objective means of classifying channel unit morphology with a correct classification rate of 85 %. An analysis of the morphological diversity along the surveyed channel indicates that reaches of at least 15 bankfull width equivalents are required to capture the channel's dominant morphological heterogeneity. Altogether, the results provide a precedent for using RPA to characterize the morphology and diversity of forested streams under dense canopies.

Structural Characterization and Comparison of Monovalent Cation-Exchanged Zeolite-W

We report comparative structural changes of potassium-contained zeolite-W (K-MER, structural analogue of natural zeolite merlinoite) and monovalent extra-framework cation (EFC)-exchanged M-MERs (M = Li+, Na+, Ag+, and Rb+). High-resolution synchrotron X-ray powder diffraction study precisely determines that crystal symmetry of MERs is tetragonal (I4/mmm). Rietveld refinement results reveal that frameworks of all MERs are geometrically composed of disordered Al/Si tetrahedra, bridged by linkage oxygen atoms. We observe a structural relationship between a group of Li-, Na-, and Ag-MER and the group of K- and Rb-MER by EFC radius and position of M(1) site inside double 8-membered ring unit (d8r). In the former group, a-axes decrease reciprocally, c-axes gradually extend by EFC size, and M(1) cations are located at the middle of the d8r. In the latter group, a- and c-axes lengths become longer and shorter, respectively, than axes of the former group, and these axial changes come from middle-to-edge migration of M(1) cations inside the d8r channel. Unit cell volumes of the Na-, Ag-, and K-MER are ca. 2005 Å3, and the volume expansion in the MER series is limited by EFC size, the number of water molecules, and the distribution of extra-framework species inside the MER channel. EFC sites of M(1) and M(2) show disordered and ordered distribution in the former group, and all EFC sites change to disordered distribution after migration of the M(1) site in the latter group. The amount of water molecules and porosities are inversely proportional to EFC size due to the limitation of volume expansion of MERs. The channel opening area of a pau composite building unit and the amount of water molecules are universally related as a function of cation size because water molecules are mainly distributed inside a pau channel.

Cooperativity and Steep Voltage Dependence in a Bacterial Channel

This paper reports on the discovery of a novel three-membrane channel unit exhibiting very steep voltage dependence and strong cooperative behavior. It was reconstituted into planar phospholipid membranes formed by the monolayer method and studied under voltage-clamp conditions. The behavior of the novel channel-former, isolated from Escherichia coli, is consistent with a linearly organized three-channel unit displaying steep voltage-gating (a minimum of 14 charges in the voltage sensor) that rivals that of channels in mammalian excitable membranes. The channels also display strong cooperativity in that closure of the first channel permits the second to close and closure of the second channel permits closure of the third. All three have virtually the same conductance and selectivity, and yet the first and third close at positive potentials whereas the second closes at negative potentials. Thus, is it likely that the second channel-former is oriented in the membrane in a direction opposite to that of the other two. This novel structure is named “triplin.” The extraordinary behavior of triplin indicates that it must have important and as yet undefined physiological roles.

Habitat-specific production of aquatic and terrestrial invertebrate drift in small forest streams: implications for drift-feeding fish

The influence of stream channel structure on the production of prey for drift-feeding fish is not well understood. We quantified drift production, the entry rate per streambed area, and drift flux, the total export rate per channel unit, in three second-order, forested streams in southwest British Columbia, Canada. We tested whether (1) drift production was higher in riffles than in pools, (2) drift flux increased with riffle length, and (3) prey body size was larger from riffles relative to pools and from terrestrial drop relative to aquatic drift. Total and aquatic-derived drift production (mg·m−2·h−1) was ∼3.5 times higher in riffles relative to pools; however, terrestrial drift did not differ between channel types. Total drift flux (mg·h−1) was positively related to riffle length. Terrestrial invertebrates were approximately three times larger than aquatics, and invertebrates from riffles were approximately three times larger on average than those from pools. These results suggest that channel structure directly affects prey availability and prey quality for drift-feeding fish and that long riffles may be key areas of prey generation.