Human-in-the-Loop Segmentation of Earth Surface Imagery

Segmentation, or the classification of pixels (grid cells) in imagery, is ubiquitously applied in the natural sciences. Manual methods are often prohibitively time-consuming, especially those images consisting of small objects and/or significant spatial heterogeneity of colors or textures. Labeling complicated regions of transition that in Earth surface imagery are represented by collections of mixed-pixels, -textures, and -spectral signatures, can be especially error-prone because it is difficult to reliably unmix, identify and delineate consistently. However, the success of supervised machine learning (ML) approaches is entirely dependent on good label data. We describe a fast, semi-automated, method for interactive segmentation of N-dimensional (x,y,N) images into two-dimensional (x,y) label images. It uses human-in-the-loop ML to achieve consensus between the labeler and a model in an iterative workflow. The technique is reproducible; the sequence of decisions made by human labeler and ML algorithms can be encoded to file, so the entire process can be played back and new outputs generated with alternative decisions and/or algorithms. We illustrate the scientific potential of segmentation of imagery of diverse settings and image types using six case studies from river, estuarine, and open coast environments. These photographic and non-photographic imagery consist of 1- and 3-bands on regular and irregular grids ranging from centimeters to tens of meters. We demonstrate high levels of agreement in label images generated by several labelers on the same imagery, and make suggestions to achieve consensus and measure uncertainty, ideal for widespread application in training supervised ML for image segmentation.

Study on Marine Water Quality Parameters in Keluang Kecil River Estuary at Bukit Keluang, Terengganu

This study presents marine water quality status of river estuarine in Keluang Kecil River Estuary at Bukit Keluang. River estuarine water is classified in Class E in MWQI. In the subsequent MWQI with 0 to 100 scales, marine water quality at 0 will be described as “Poor” and 100 as “Excellent”. This study aims to determine water quality parameters and the water quality status of river estuarine by using Malaysia Marine Water Quality Index (MMWQI). In-situ and ex-situ analyses in this study were conducted for five sampling points that had been selected in the Keluang Kecil river estuary. In-situ parameters, dissolved oxygen (DO), pH, temperature, salinity and ammonia, were measured using YSI Professional Plus Series Multi-parameter during sampling. At the same time, water samples for total suspended solids, phosphate, nitrate and faecal coliform were collected and analysed in the laboratory using a standard procedure based on American Public Health Association (APHA), 2017 and HACH methods and then were compared with the Malaysia Marine Water Quality Standard (MMWQS). The average value of ammonia, nitrate, phosphate and TSS were not within the acceptable limit or standard value set by Department of Environment (DOE) except for dissolved oxygen and faecal coliform. Based on the one-way analysis of variance (ANOVA) result, there are significant differences (P < 0.05) in DO, pH, temperature, salinity, ammonia, TSS and FC in water between stations except for nitrate and phosphate. Marine Water Quality Index (MWQI) was measured for all the sampling stations (S1, S2, S3, S4 and S5) using a formula set by DOE (2019). The mean of MWQI for the Keluang Kecil river estuary measured was 43.45. Thus, the river estuary is classified as Poor (0 – 49). This study can maintain the marine water quality to conserve estuary biodiversity. This study was needed to prevent the possibility of water pollution and water quality can be monitored in that area in the future. The understanding of this study about water quality is to maintain the water quality and facilitate the management as actions can be taken by local authorities and other government agencies to maintain and improve the water quality and create boundaries and regulations that can bring back nature.

Carbon dynamics at the river-estuarine transition: a comparison among tributaries of Chesapeake Bay

Sources and transformation of C were quantified using mass balance and ecosystem metabolism data for the upper segments of the James, Pamunkey and Mattaponi Estuaries. The goal was to assess the role of external (river inputs &amp; tidal exchange) vs. internal (metabolism) drivers in influencing the forms and fluxes of C. C forms and their response to river discharge differed among the estuaries based on their physiographic setting. The James, which receives the bulk of inputs from upland areas (Piedmont and Mountain), exhibited a higher ratio of inorganic to organic C, and larger inputs of POC. The Pamunkey and Mattaponi receive a greater proportion of inputs from lowland (Coastal Plain) areas, which were characterized by low DIC and POC, and elevated DOC. We anticipated that transport processes would dominate during colder months when discharge is elevated and metabolism is low, and that biological processes would predominate in summer, leading to attenuation of C through-puts via de-gassing of CO2. Contrary to expectations, highest retention of OC occurred during periods of high through-put, as elevated discharge resulted in greater loading and retention of POC. In summer, internal cycling of C via production and respiration was large in comparison to external forcing despite the large riverine influence in these upper estuarine segments. The estuaries were found to be net heterotrophic based on retention of OC, export of DIC, low GPP relative to ER, and a net flux of CO2 to the atmosphere. In the James, greater contributions from phytoplankton production resulted in a closer balance between GPP and ER, with autochthonous production exceeding allochthonous inputs. Combining the mass balance and metabolism data with bioenergetics provided a basis for estimating the proportion of C inputs utilized by the dominant metazoan. The findings suggest that invasive catfish utilize 15 % of total OM inputs and up to 40 % of allochthonous inputs to the James.

Assessing the Rigidity of the River Continuum Concept to Fish Species in a Fluvially Dominated Southern African Coastal System

A continuous gradient of physical and biological processes exists within a river system from headwaters to mouth, according to the River Continuum Concept. In this study, the River Continuum Concept was used to compare patterns in fish assemblages and diversity trends to Remane predictions. The findings show that species diversity is low near the estuary's mouth and increases upstream. While the presence of several marine and estuarine species increased species diversity upstream in the river due to the dominance of freshwater fishes, it decreased downstream due to the presence of several marine and estuarine species. These patterns are consistent with the biogeographical trend of decreasing species richness along the South African coast from east to west. Based on the findings of this study, the River Continuum Concept ignores plasticity, which is evident in species that use transitional waters. As a result, the concept is insufficiently valid for the Orange River Estuarine Continuum, and a new concept known as the “River-Estuarine Continuum” has been proposed.