No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Towards a framework for terrain attribute selection in environmental studies
Abstract
Terrain attributes (e.g. slope, rugosity) derived from digital terrain models are commonly used in environmental studies. The increasing availability of GIS tools that generate those attributes can lead users to select a sub-optimal combination of terrain attributes for their applications. Our objectives were to identify sets of terrain attributes that best capture terrain properties and to assess how they vary with surface complexity. 230 tools from 11 software packages were used to derive terrain attributes from nine surfaces of different topographic complexity levels. Covariation and independence of terrain attributes were explored using three multivariate statistical methods. Distinct groups of correlated terrain attributes were identified, and their importance in describing a surface varied with surface complexity. Terrain attributes were highly covarying and sometimes ambiguously defined within software documentation. We found that a combination of six to seven particular terrain attributes always captures more than 70% of the topographic structure of surfaces.
Supplementary resource (1)
Data
November 2017
... Terrain attributes that describe the shape and character of the Earth's surface can be calculated from these DTMs. While a large number of attributes can be found in the literature (Wilson, 2018), the most common terrain attributes can generally be categorized into five thematic groups: slope, aspect, curvature, relative position, and roughness (Bouchet et al., 2015;Lecours et al., 2016Lecours et al., , 2017Wilson et al., 2007) (Figure 1). While this framework is by no means the only way to group terrain attributes, it provides a useful way to organize related terrain attributes based on their physical meaning and will be used within this article for organizational purposes. ...
... Although widely used, the selection and application of terrain attributes to a particular task may often be suboptimal (Ironside et al., 2018;Lecours et al., 2017). Owing to the abundance of attributes found in the literature and measures being dispersed across various software packages, users may not fully understand how the terrain attributes they select are calculated, and may only choose a subset of readily available attributes, which can result in a failure to fully capture useful information that could be derived from the DTM (Bouchet et al., 2015;Lecours et al., 2017). ...
... Although widely used, the selection and application of terrain attributes to a particular task may often be suboptimal (Ironside et al., 2018;Lecours et al., 2017). Owing to the abundance of attributes found in the literature and measures being dispersed across various software packages, users may not fully understand how the terrain attributes they select are calculated, and may only choose a subset of readily available attributes, which can result in a failure to fully capture useful information that could be derived from the DTM (Bouchet et al., 2015;Lecours et al., 2017). For example, fields such as ecology often use terrain attributes as predictors for species distribution models (Franklin, 1995;Ironside et al., 2018;Mod et al., 2016). ...
Digital terrain models (DTMs) are datasets containing altitude values above or below a reference level, such as a reference ellipsoid or a tidal datum over geographic space, often in the form of a regularly gridded raster. They can be used to calculate terrain attributes that describe the shape and characteristics of topographic surfaces. Calculating these terrain attributes often requires multiple software packages that can be expensive and specialized. We have created a free, open-source R package, MultiscaleDTM, that allows for the calculation of members from each of the five major thematic groups of terrain attributes: slope, aspect, curvature, relative position, and roughness, from a regularly gridded DTM. Furthermore, these attributes can be calculated at multiple spatial scales of analysis, a key feature that is missing from many other packages. Here, we demonstrate the functionality of the package and provide a simulation exploring the relationship between slope and roughness. When roughness measures do not account for slope, these attributes exhibit a strong positive correlation. To minimize this correlation, we propose a new roughness measure called adjusted standard deviation. In most scenarios tested, this measure produced the lowest rank correlation with slope out of all the roughness measures tested. Lastly, the simulation shows that some existing roughness measures from the literature that are supposed to be independent of slope can actually exhibit a strong inverse relationship with the slope in some cases.
... Soil factors generally influence the temporal and spatial variance of R s rate, e.g., soil organic carbon content (SOC), soil physical and chemical properties (texture, bulk density, pH, etc.). The density and activity of plant roots through R aut as well as the activity of soil microbial biomass through R het have great influence on R s rate too (Lellei-Kovács et al., 2016;Michelsen et al., 2004;Mitra et al., 2019;Thomas et al., 2018). The share of R s components is largely dependent on these variables (Hu et al., 2001;Moyes and Bowling, 2016;Raich and Tufekcioglu, 2000;Tang et al., 2020), responding to the changes of environmental factors differently (Balogh et al., 2016;Shi et al., 2011). ...
... The share of R s components is largely dependent on these variables (Hu et al., 2001;Moyes and Bowling, 2016;Raich and Tufekcioglu, 2000;Tang et al., 2020), responding to the changes of environmental factors differently (Balogh et al., 2016;Shi et al., 2011). Among the soil parameters, R s shows a positive temporal relationship with soil temperature (T s ) and soil water content (SWC) (Balogh et al., 2015;Lellei-Kovács et al., 2016;Moyes and Bowling, 2016;Shi et al., 2011;Tang et al., 2020). Thus, seasonal changes in soil parameters play a significant role in defining seasonal differences in soil CO 2 emission (Hao et al., 2010;Herbst et al., 2009;Mitra et al., 2019;Raich and Tufekcioglu, 2000;Thomas et al., 2018). ...
... where z i values are the elevations of the correspondent R radius while z is the mean elevation within R. SD describes the heterogeneity and local surface roughness within the raster. For slope-and aspect-derived easterness and northness, we used eight neighbors, as suggested by Lecours et al. (2017): ...
Forest–steppe habitats in central Hungary have contrasting canopy structure
with strong influence on the spatiotemporal variability of ecosystem
functions. Canopy differences also co-vary with terrain feature effects,
hampering the detection of key drivers of carbon cycling in this threatened
habitat. We carried out seasonal measurements of ecosystem functions (soil
respiration and leaf area index), microclimate and soil variables as well
as terrain features along transects for 3 years in poplar groves and the
surrounding grasslands. We found that the terrain features and the canopy
differences co-varyingly affected the abiotic and biotic factors of this
habitat. Topography had an effect on the spatial distribution of soil
organic carbon content. Canopy structure had a strong modifying effect
through allocation patterns and microclimatic conditions, both affecting
soil respiration rates. Due to the vegetation structure difference between
the groves and grasslands, spatial functional diversity was observed. We
found notably different conditions under the groves with high soil
respiration, soil water content and leaf area index; in contrast, on the
grasslands (especially in E–SE–S directions from the trees) soil temperature
and vapor pressure deficit showed high values. Processes of aridification
due to climate change threaten these habitats and may cause reduction in the
amount and extent of forest patches and decrease in landscape diversity.
Owing to habitat loss, reduction in carbon stock may occur, which in turn
has a significant impact on the local and global carbon cycles.
... All predictors were processed to ensure a common resolution of 250 m  250 m, at depths between 1 and 50 m. ARCGIS 10.5.1 was used for all processing, with terrain derivatives created using TERRAIN ATTRIBUTE SELECTION FOR SPATIAL ECOLOGY 1.1 (TASSE; Lecours et al., 2017) and BENTHIC TERRAIN MODELER 3.0 (BTM; Walbridge et al., 2018) toolboxes. ...
... To maintain model simplicity and avoid overfitting, indicators of primary productivity were included as a composite variable. As recommended by Lecours et al. (2017), terrain attributes were derived from the digital bathymetric model and included in the analysis. A full list of variables considered can be found in Table S1, whereas greater detail on variable consideration and rationale can be found in Table S2. ...
... Spearman's rank correlation coefficients and significance tests were then applied to data extracted from the remaining predictors with 1,000 random points (Stirling et al., 2016;Lecours et al., 2017). ...
• As an increasingly important resource in ecological research, citizen scientists have proven dynamic and cost-effective in the supply of data for use within habitat suitability models. With predictions critical to the provision of effective conservation measures in cryptic marine species, this study delivers baseline ecological data for the Critically Endangered angelshark (Squatina squatina), exploring: (i) seasonal, sex-differentiated distributions; (ii) environmental distribution predictors; and (iii) examining bias-corrected, imperfect citizen science data for use in coastal habitat suitability models with cryptic species.
• Citizen science presence data, comprising over 60,000 hours of sampling effort, were used alongside carefully selected open-source predictor variables, with maxent generating seasonal male and female habitat suitability models for angelsharks in the Canary Islands. A biased prior method was used, alongside two model validation measures to ensure reliability.
• Citizen science data used within maxent suggest that angelshark habitat suitability is low in coastal areas during warmer months, with fewer occurrences despite a negligible change in sampling effort. The prime importance of bathymetry may indicate the importance of depth for reproductive activity and possible diel vertical migration, whereas aspect may act as a proxy for sheltered habitats away from open ocean. Substrate as a predictor of female habitats in spring and summer could imply that soft sediment is sought for birthing areas, assisting in the identification of areas critical to reproductive activity and thus locations that may benefit from spatial protections.
• Model outputs to inform recovery plan development and ecotourism are identified as plausible safeguards of population recovery, whereas the comparison of biased and bias-corrected models highlights some variance between methodologies, with bias-corrected models producing greater areas of habitat suitability. Accordingly, an adaptive framework is provided for the implementation of citizen science data within the modelling of cryptic coastal species distribution.
... TAs as well as DEMs are represented in raster format, which deliver information cell by cell (or pixel by pixel) in a continuous way (Arrouays et al. 2017). As a consequence, the spatial resolution may influence the results (Lecours et al. 2017;Mokarram and Hojati 2017;Penížek et al. 2016). Although fine resolution provides greater topographic detail, it can also introduce local artifacts and slow down data processing, whereas coarse resolution furnishes less detail and often provides missing channels and peaks (Hengl 2006). ...
... All of the models used TA data and PM. In contrast to TA data, which are easily obtained worldwide through DEM at diverse spatial resolutions Gray et al. 2016;Hengl et al. 2017;Heung et al. 2016;Lecours et al. 2017;Shangguan et al. 2014;Silva et al. 2016a), detailed PM information at large scales is scarce in many parts of the world, especially in developing countries such as Brazil. For this study, a PM map of the study area was obtained from Curi et al. (2017) who employed magnetic susceptibility data of soils using fuzzy logic. ...
... In digital soil mapping, resolution has an inverse relationship with the size of the area represented by the basic mapping unit, the pixel. Also, the increase or decrease in the resolution directly implies the values extracted from the different terrain attributes derived from the DEM (Mashimbye et al. 2014) and can still vary with the degree of surface complexity (Lecours et al. 2017). Terrain attributes whose values are computed by accumulating adjacent pixel values (e.g., flow accumulation) have their range of values increased in smaller resolutions, whereas morphometric attributes (e.g., slopes, curvatures) decrease the range of values (Penížek et al. 2016). ...
In developing countries, the use of proximal and remotely sensed data is of critical importance as a less expensive means of obtaining soils information. While proximal sensor approaches such as portable X-ray fluorescence (pXRF) spectrometry are becoming increasingly used to predict soil properties worldwide, remotely sensed data has also been used for terrain analysis in recent decades with the aid of powerful interpretive algorithms. The aims of this work were to apply a random forest algorithm to model and predict the available contents of Fe, Cu, Mn, and Zn from pXRF data in addition to terrain attributes (TAs) with 5 and 10 m spatial resolution and parent material information. The data were used separately and together in an area with high variability of parent materials. Soil samples (n = 153) were collected, analyzed by pXRF, and subjected to laboratory analyses to determine the available contents of Fe, Cu, Mn, and Zn. Twelve TAs were generated from digital elevation models (DEM). These data were divided into five datasets (or random forest inputs): pXRF data; TA 5 m data; TA 10 m; pXRF + TA 5 m; and pXRF + TA 10 m. Predictions were performed to assess the importance of such variables. Models were validated with an independent set of samples. Finally, the best models were spatially rendered to cover the entire study area and maps were also validated. The combination of pXRF data and TA covariates in addition to parent material information allowed accurate predictions of available Fe, Mn, Cu, and Zn through the random forest algorithm. Parent material information improved the predictions. Pixel size of 10 m resolution promoted better results than 5 m resolution. Available Fe contents were better predicted using only TA data. For the spatial prediction of available micronutrients, validation of maps resulted in R² of 0.88, RMSE of 59.97 mg kg⁻¹ and ME of 24.00 mg kg⁻¹ for Fe; 0.85, 29.65 mg kg⁻¹, 9.70 mg kg⁻¹ for Mn, 0.64, 3.11 mg kg⁻¹, 0.71 mg kg⁻¹ for Zn and 0.82, 1.17 mg kg⁻¹, 0.43 mg kg⁻¹ for Cu, respectively. Available micronutrient contents can be accurately predicted using pXRF data in association with terrain and parent material information.
... In addition, ground-truthing data, for example, collected from camera imagery, must be collected in order to inform or verify interpretations of the acoustic returns and understand what sediment and other geological features they represent and to identify the biota associated with these habitats [7,8,17]. Additionally, with the increasing volume of data and the greater demand for habitat maps for ecosystem-based management and marine spatial planning, scientists and managers must consider the reliability of MBES data and their derivative metrics as surrogates to classify habitat types over differing spatial scales, especially where no independent ground-truthing exists [9,[17][18][19][20][21]. For bathymetry, these derivative features include terrain attributes such as slope, rugosity, aspect, etc. [21,22], and for backscatter, these include texture metrics such as those derived from a gray level co-occurrence matrix (GLCM) [23][24][25][26][27][28][29]. ...
... Additionally, with the increasing volume of data and the greater demand for habitat maps for ecosystem-based management and marine spatial planning, scientists and managers must consider the reliability of MBES data and their derivative metrics as surrogates to classify habitat types over differing spatial scales, especially where no independent ground-truthing exists [9,[17][18][19][20][21]. For bathymetry, these derivative features include terrain attributes such as slope, rugosity, aspect, etc. [21,22], and for backscatter, these include texture metrics such as those derived from a gray level co-occurrence matrix (GLCM) [23][24][25][26][27][28][29]. ...
... The 10 × 10 m resolution bathymetry and backscatter grids were used to calculate various derivative metrics (Table 1). These metrics include terrain attributes derived from the bathymetry surface [21,22] and texture measures derived from the backscatter mosaic using GLCMs [23,24] as well as the local mean and the standard deviation of backscatter using varying window sizes. The 10 × 10 m bathymetry and backscatter surfaces along with their derivative features were used as the predictor variables in the statistical models. ...
The west Florida shelf (WFS; Gulf of Mexico, USA) is an important area for commercial and recreational fishing, yet much of it remains unmapped and unexplored, hindering effective monitoring of fish stocks. The goals of this study were to map the habitat at an intensively fished area on the WFS known as “The Elbow”, assess the differences in fish communities among different habitat types, and estimate the abundance of each fish taxa within the study area. High-resolution multibeam bathymetric and backscatter data were combined with high-definition (HD) video data collected from a near-bottom towed vehicle to characterize benthic habitat as well as identify and enumerate fishes. Two semi-automated statistical classifiers were implemented for obtaining substrate maps. The supervised classification (random forest) performed significantly better (p = 0.001; α = 0.05) than the unsupervised classification (k-means clustering). Additionally, we found it was important to include predictors at a range of spatial scales. Significant differences were found in the fish community composition among the different habitat types, with both substrate and vertical relief found to be important with rock substrate and higher relief areas generally associated with greater fish density. Our results are consistent with the idea that offshore hard-bottom habitats, particularly those of higher vertical relief, serve as “essential fish habitat”, as these rocky habitats account for just 4% of the study area but 65% of the estimated total fish abundance. However, sand contributes 35% to total fish abundance despite comparably low densities due to its large area, indicating the importance of including these habitats in estimates of abundance as well. This work demonstrates the utility of combining towed underwater video sampling and multibeam echosounder maps for habitat mapping and estimation of fish abundance.
... Commercial and freeware terrain analysis software has proliferated in stand-alone, image analysis or Geographical Information System (GIS) software packages, such as ArcGeomorphometry (Rigol-Sanchez, Stuart, and Pulido-Bosch 2015), Whitebox Geospatial Analysis Tools (GAT) (Lindsay 2016), System for Automated Geoscientific Analyses (SAGA) (Conrad et al. 2015), Topographic Analysis Toolkit (TAK) (Forte and Whipple 2019), and Geomorphometry and Gradient Metrics Toolbox (GGMT) (Evans et al. 2014). Many geomorphometric processing tools are also applicable to Digital Surface Models (DSMs) (Hirt 2014) and diverse remote sensing image datasets (Middleton et al. 2015;Diesing, Mitchell, and Stephens 2016;Lecours et al. 2017). ...
... Unfortunately, it is often not known a priori which geomorphometric variables control a given process at any given location (Florinsky 2017). Careful variable selection based on previous work, expert-input, field testing, and statistical tools, such as variable intercorrelations, rankings, and Principal Components Analysis (PCA), is recommended (Lecours et al. 2017). This critical step in the use of geomorphometry in remote sensing applications is a theme throughout this paper. ...
... The dominant landform, hydrological and ecological processes and equilibrium states are scale-dependent and will influence the appropriate DEM spatial detail and the variables selected for any given application. Operationally, there are competing pressures; for example, between variable redundancy, which may contribute to suboptimal variable selection (Lecours et al. 2017), and the importance of subtle topographic effects, which may not be statistically significant but are substantive in interpretation. Finally, many of these studies have suggested that further methodological advances are possible with greater attention to issues involved in the variable definition, derivation, and integration. ...
An understanding of topographic surfaces and a system of geomorphometric variables underlies the effective use of Digital Elevation Models (DEMs) in remote sensing. This paper is focussed on the basic concepts and structure of geomorphometric variable types relevant to geophysical and biophysical remote sensing applications. In general, remote sensing analysts must be satisfied that there is a reasonable expectation that geomorphometric processing will benefit the remote sensing project at hand based on physical or mathematical relationships. Careful selection of local, textural, and contextual geomorphometric variables is required. The characteristics of these data can help optimize the analysis approach, which increasingly involves machine learning algorithms and Object-based DEM Analysis (OBDA) methods. Examples from the literature and a DEM of the Peterborough Drumlin Field are used to illustrate the discussion.
... No terrain models can really be free of artifacts (Lecours et al. 2017a). The habitats are represented on maps as clusters of pixels sharing some characteristics with model data (Lecours et al. 2017b). Thus, it must be considered that the artifacts influence the boundaries of the benthic structures (Lecours et al. 2017b), influencing the final classification. ...
... The habitats are represented on maps as clusters of pixels sharing some characteristics with model data (Lecours et al. 2017b). Thus, it must be considered that the artifacts influence the boundaries of the benthic structures (Lecours et al. 2017b), influencing the final classification. In our study, we recognize that the bathymetric data sets were compiled from nautical charts, resulting in anisotropic features from the spatial interpolation process of these data which may influence the terrain attributes and the final classification of geomorphic features (see Lecours et al. 2017b;Goes et al. 2019). ...
... Thus, it must be considered that the artifacts influence the boundaries of the benthic structures (Lecours et al. 2017b), influencing the final classification. In our study, we recognize that the bathymetric data sets were compiled from nautical charts, resulting in anisotropic features from the spatial interpolation process of these data which may influence the terrain attributes and the final classification of geomorphic features (see Lecours et al. 2017b;Goes et al. 2019). Residual effects of the interpolation are found in the southern of area of the PCS where bathymetric data points were at a lower spatial density, which resulted in the presence of the uncertain crests and depressions. ...
Marine habitat mapping provides essential information for environmental management and design of marine reserves. In tropical regions, particularly along the Brazilian coast, the spatial variability of marine habitats is poorly known. The aim of this study is to evaluate the geodiversity of the region as an indicator to benthic habitat distribution, applying an integrative approach utilizing existing broad-scale bathymetric and seafloor geological data sets. A digital bathymetric model (DBM) of the Pernambuco Continental shelf (PCS) was generated at 80 m resolution from available bathymetric data. Through the benthic terrain model (BTM), DBM derivatives and the benthic structures were generated. These structures were combined with the textural seabed classification using tools in ArcGIS™ 10.5 Spatial Analyst to identify 22 seabed geomorphic features, including a submarine canyon confirmed around 8°20′S. These geomorphological features describe the surficial characteristics of the seafloor, providing the baseline for subsequent habitat-mapping studies, and can therefore be considered a potential habitat map. Three profiles taken to describe the cross-shelf seafloor environment (North, Center, and South) extracted from interpolated seabed geomorphological map revealed that sediment grain size becomes coarser from inshore to offshore, with predominantly sand and gravel sediment grain sizes. Overall, these results indicated a great potential for PCS due to the geodiversity present in the area, which leads to a higher biodiversity as well. Based on that, the central portion of PCS is suggested as a priority area for conservation because it was the most geomorphologically diverse. These terrain features influence the environmental conditions, such as currents, waves, nutrients, and other oceanographic parameters, which results in high diversity of benthic habitats in the region. This study presents a first step in characterizing the geomorphology of PCS by using reliable, standardized data.
... Multiple geomorphometric parameters have been developed and used to characterise the terrain variability (Hengl, Reuter, 2009;Wilson, Gallant, 2000). At the same time, Lecours et al. (2017) argued that more than 70% of environmental variability related to the relief is captured through the limited set of the six parameters, namely, local mean, slope, local standard deviation, relative difference to mean, easterness and northerness. I derived the suggested parameters from the AW3D30 DSM in the SAGA GIS (Conrad et al., 2015) that provides the algorithms recommended by Lecours et al. (2017). ...
... At the same time, Lecours et al. (2017) argued that more than 70% of environmental variability related to the relief is captured through the limited set of the six parameters, namely, local mean, slope, local standard deviation, relative difference to mean, easterness and northerness. I derived the suggested parameters from the AW3D30 DSM in the SAGA GIS (Conrad et al., 2015) that provides the algorithms recommended by Lecours et al. (2017). ...
... Covariation is a common property of many environmental variables, especially of those derived from a single data source such as digital elevation model (DEM) (Graham, 2003;Lecours et al., 2017). To avoid possible data redundancy, I performed the correlation analysis to account for confounding spatial variables (Dutilleul et al., 1993;Osorio, Vallejos, 2014;R Core Team, 2017). ...
As anisotropy is a fundamental property of the real-world environmental spatial variables, the conventional omnidirectional variograms and correlograms do not provide means enough to characterise spatial dependence between observations. The purpose of this article is to introduce directional univariate structure functions analysis to explore and quantify the spatial anisotropy of environmental variables. Analysis of six environmental variables within three physical–geographical regions proved the leading role of relief for landscape differentiation; it also defined the size and extension of major landforms responsible for the organisation of spatial pattern. The arrangement of the vegetation patches demonstrated linkage with the major landforms. The other relief derivatives, being prone to noise and artefacts in the original data, showed a random-variable type of behaviour. In the lack of any particular spatially anisotropic structure, the results of the analysis can provide a clue about meaningful distances of interest at finer scales. The approach can also be an exploratory tool for discrete measurements to recognise the features of spatial continuity.
... Geomorphometry is a quantitative earth surface analytical science that originated from geomorphology and evolved from mathematics, earth sciences, and computer science [33,38]. Marine geomorphometry refers to the generation of terrain attributes (e.g., slope, aspect, and curvature) from bathymetry which is known as general geomorphometry, and the extraction of discrete seabed features (e.g., valleys, seamounts, and ridges) from bathymetry which is known as specific geomorphometry [39][40][41]. ...
... The terrain attributes that are known to capture most of the seafloor general geomorphologic characteristics are broad-and fine-scales Bathymetric Position Index (BPI), standard deviation, rugosity, aspect, bathymetric mean, curvature, ruggedness, and slope [33,39,40,[42][43][44][45][46][47][48][49][50]. The slope, broad and fine scales BPI, and terrain ruggedness/roughness were found to be suitable for the analysis of this study ( Table 1). ...
The Caroline Ridge (CR) subduction underneath the Philippine Sea Plate brings complex morphotectonic characteristics to the Yap Subduction Zone (YSZ) compared to other normal intra-oceanic subduction systems. However, due to the relative paucity of precise geomorphological information, the detailed morphotectonic settings of the YSZ remain unclear. Therefore, we combine the latest-released bathymetry, marine geomorphometry techniques, and geophysical information to investigate the geomorphological characteristics of landforms in the YSZ and their inter-relationship with the CR subduction. The Parece Vela Basin displays NE-SW oriented fractures which are believed to be influenced by the subduction of CR in the ESE-WNW direction. The north part of the Yap arc exhibits higher Bouguer anomalies, implying the absence of the overlying normal volcanic arc crust. The arc-ward trench shows abnormal higher slope values and reveals two significant slope breaks. The Yap Trench axis reveals varying water depths with an extraordinarily deep point at around 9000 m. The sea-ward trench slope displays higher slope values than normal and shows the presence of grabens, horsts, and normal faults which indicate the bending of the CR before subduction. The CR subduction is observed to be critical in the formation of significant geomorphological characteristics in the YSZ.
... They include seabed slope to measure the surface gradient, and northness and eastness to assess its orientation (Wilson et al., 2007). Relative difference to the mean value (RDMV) measures topographic position and bathymetric standard deviation (SD) reflects seabed roughness (Walbridge et al., 2018;Lecours et al., 2017). A comparable metric, 'vector ruggedness measure' (VRM) is also frequently used to measure seabed roughness, however, Lecours et al. (2017) found that both measures were correlated and recommended SD to measure rugosity. ...
... Relative difference to the mean value (RDMV) measures topographic position and bathymetric standard deviation (SD) reflects seabed roughness (Walbridge et al., 2018;Lecours et al., 2017). A comparable metric, 'vector ruggedness measure' (VRM) is also frequently used to measure seabed roughness, however, Lecours et al. (2017) found that both measures were correlated and recommended SD to measure rugosity. Additionally, measures of curvature (curvature, profile, and planar curvature) can highlight contours such as ridges, valleys, and mounds (Walbridge et al., 2018;Wilson et al., 2007). ...
Seafloor habitat maps are an important management tool used to delineate distinct regions of the seabed based on their biophysical properties. Spatially continuous bathymetry and backscatter-derived terrain features are commonly used as proxies for environmental conditions and processes that affect species distributions. Multi-scale approaches are increasingly applied to assess the relevant scales at which species co-occur. As the optimal scale(s) may be unknown, features can be calculated at multiple successive scales, yet this results in numerous highly correlated features that may negatively impact model interpretability. To address this increased dimensionality, feature selection approaches can be used to identify the most relevant features. Here, filter and wrapper approaches are assessed to select features from a highly dimensional multi-scale dataset. Terrain features describing the seabed were calculated across ten scales at two coastal sites in Placentia Bay, Newfoundland, Canada. Five species assemblages were identified using ground-truth underwater video sampling. Features predicting the presence of assemblages were assessed using the two selection methods, and the set of chosen features was modelled using three machine learning algorithms: extreme gradient boosting (XGB), random forest (RF), and support vector machines (SVM). The XGB model with features selected by scale-factor from the Boruta wrapper algorithm had the highest accuracy according to cross-validation- (61.67%, kappa 0.49). Bathymetry and terrain attributes were the most important predictors of assemblage occurrence across various analysis scales encompassing both broader and fine-scale variability of the seabed. The proposed feature reduction and selection approach improved the overall accuracy of predictions, and the resulting biological complexity captured in our habitat maps established baseline data for an ecologically significant coastal region.
... We generated six morphological derivatives from the bathymetry data using the Surface Parameters and Raster Calculator tools in ArcGIS Pro (ESRI, CA, USA) and the Benthic Terrain Modeller v3.0 plugin (Walbridge et al., 2018;Wright et al., 2005). The derivatives we used were slope, curvature, eastness, northness, relative difference from mean value (RDMV) and vector ruggedness measure (VRM) (Lecours et al., 2017;Sappington et al., 2007;Wilson et al., 2007). We selected these based on their demonstrated predictive power in the literature, their hypothesised predictive power within the context of this study, and following recommendations from Lecours et al. (2017). ...
... The derivatives we used were slope, curvature, eastness, northness, relative difference from mean value (RDMV) and vector ruggedness measure (VRM) (Lecours et al., 2017;Sappington et al., 2007;Wilson et al., 2007). We selected these based on their demonstrated predictive power in the literature, their hypothesised predictive power within the context of this study, and following recommendations from Lecours et al. (2017). Morphological derivatives are typically calculated using a square window with an edge length of 3 pixels, but the scale at which they are generated and the way in which they are calculated for different scales can influence their predictive power (Misiuk et al., 2021;Porskamp et al., 2018). ...
High energy marine regions host ecologically important habitats like temperate reefs, but are less anthropogenically developed and understudied compared to lower energy waters. In the marine environment direct habitat observation is limited to small spatial scales, and high energy waters present additional logistical challenges and constraints. Semi-automated predictive habitat mapping is a cost-effective tool to map benthic habitats across large extents, but performance is context specific. High resolution environmental data used for predictive mapping are often limited to bathymetry, acoustic backscatter and their derivatives. However, hydrodynamic energy at the seabed is a critical habitat structuring factor and likely an important, yet rarely incorporated, predictor of habitat composition and spatial patterning. Here, we used a machine learning classification approach to map temperate reef substrate and biogenic reef habitat in a tidal energy development area, incorporating bathymetric derivatives at multiple scales and simulated tidally induced seabed shear stress. We mapped reef substrate (four classes: sediment (not reef), stony reef (low resemblance), stony reef (medium – high resemblance) and bedrock reef) with overall balanced accuracy of 71.7%. Our model to predict potential biogenic Sabellaria spinulosa reef performed less well with an overall balanced accuracy of 63.4%. Despite low performance metrics for the target class of potential reef in this model, it still provided insight into the importance of different environmental variables for mapping S. spinulosa biogenic reef habitat. Tidally induced mean bed shear stress was one of the most important predictor variables for both reef substrate and biogenic reef models, with ruggedness calculated at multiple scales from 3 m to 140 m also important for the reef substrate model. We identified previously unresolved relationships between temperate reef spatial distribution, hydrodynamic energy and seabed three-dimensional structure in energetic waters. Our findings contribute to a better understanding of the spatial ecology of high energy marine ecosystems and will inform evidence-based decision making for sustainable development, particularly within the growing tidal energy sector.
... As the variable selected can alter the outputs of ecological models, they must be carefully selected [64]. Indeed, [13] have compiled a table that groups frequently covarying variables together so that the reader can select six to seven terrain attributes that likely capture about 70% of surface structures across a site. We recommend that the reader consults this table prior to selecting variables for their own research. ...
... When selecting variables for ecological models, one must first ensure independence between variables to avoid redundancy [13,19]. One option is to reduce the dimensionality of the data by performing a principal component analysis (PCA) with all variables that may be relevant to the study, then use the coordinates of the PCA-components as uncorrelated input predictor variables in the model. ...
The vulnerability of alpine environments to climate change presses an urgent need to
accurately model and understand these ecosystems. Popularity in the use of digital elevation models (DEMs) to derive proxy environmental variables has increased over the past decade, particularly as DEMs are relatively cheaply acquired at very high resolutions (VHR; <1 m spatial resolution).
Here, we implement a multiscale framework and compare DEM-derived variables produced by Light Detection and Ranging (LiDAR) and stereo-photogrammetry (PHOTO) methods, with the aim of assessing their relevance and utility in species distribution modelling (SDM). Using a case study on the arctic-alpine plant, Arabis alpina, in two valleys in the western Swiss Alps, we show that both LiDAR and PHOTO technologies can be relevant for producing DEM-derived variables for use in SDMs. We demonstrate that PHOTO DEMs, up to a spatial resolution of at least 1m, rivalled the accuracy of LiDAR DEMs, largely owing to the customizability of PHOTO DEMs to the study sites compared to commercially available LiDAR DEMs. We obtained DEMs at spatial resolutions of
6.25 cm–8 m for PHOTO and 50 cm–32 m for LiDAR, where we determined that the optimal spatial resolutions of DEM-derived variables in SDM were between 1 and 32 m, depending on the variable and site characteristics. We found that the reduced extent of PHOTO DEMs altered the calculations of all derived variables, which had particular consequences on their relevance at the site with heterogenous terrain. However, for the homogenous site, SDMs based on PHOTO-derived variables generally had higher predictive powers than those derived from LiDAR at matching resolutions.
From our results, we recommend carefully considering the required DEM extent to produce relevant derived variables. We also advocate implementing a multiscale framework to appropriately assess the ecological relevance of derived variables, where we caution against the use of VHR-DEMs finer than 50 cm in such studies.
... As the variable selected can alter the outputs of ecological models, they must be carefully selected [63]. Indeed, [13] have compiled a table that groups frequently covarying variables together so that the reader can select six to seven terrain attributes that likely capture about 70% of surface structures across a site. We recommend that the reader consults this table prior to selecting variables for their own research. ...
... When selecting variables for ecological models, one must first ensure independence between variables to avoid redundancy [13,19]. One option is to reduce the dimensionality of the data by performing a principle component analysis (PCA) with all variables that may be relevant to the study, then use the coordinates of the PCA-components as uncorrelated input predictor variables in the model. ...
The vulnerability of alpine environments to climate change presses an urgent need to accurately model and understand these ecosystems. Popularity in use of digital elevation models (DEMs) to derive proxy environmental variables has increased over the past decade, particularly as DEMs are relatively cheaply acquired at very high resolutions (VHR; <1m spatial resolution). Here, we implement a multiscale framework and compare DEM-derived variables produced by Light Detection and Ranging (LiDAR) and stereo-photogrammetry (PHOTO) methods, with the aims of assessing their relevance and utility in species distribution modelling (SDM). Using a case study on the arctic-alpine plant Arabis alpina in two valleys in the western Swiss Alps, we show that both LiDAR and PHOTO technologies can be relevant for producing DEM-derived variables for use in SDMs. We demonstrate that PHOTO DEMs rivalled the accuracy of LiDAR, putting the current paradigm of LiDAR being the more accurate of the two methods into question. We obtained DEMs at spatial resolutions of 6.25cm-8m for PHOTO and 50cm-32m for LiDAR, where we determined that the optimal spatial resolutions of DEM-derived variables in SDM were between 1 and 32m, depending on the variable and site characteristics. We found that the reduced extent of PHOTO DEMs altered the calculations of all derived variables, which had particular consequences on their relevance at the site with heterogenous terrain. However, for the homogenous site, we found that SDMs based on PHOTO-derived variables generally had higher predictive powers than those derived from LiDAR at matching resolutions. From our results, we recommend carefully considering the required DEM extent to produce relevant derived variables. We also advocate implementing a multiscale framework to appropriately assess the ecological relevance of derived variables, where we caution against the use of VHR-DEMs finer than 50cm in such studies.
... A suite of nine geomorphometrics was derived from the multi-beam bathymetry grids (Supplementary Figure S3), and chosen to capture a range of meaningful terrain attributes . Among them were indices of seabed aspect, slope, and curvature, as well as several measures of terrain variability retained to encapsulate independent facets of topographic complexity, and calculated on both small (100 m) and large (1000 m) scales (Lecours et al., 2017; Supplementary Table S3). The topographic position index (TPI) was preferred over the relative difference to mean value (Lecours et al., 2017) as it could be more easily calculated at different spatial scales, given available software. ...
... Among them were indices of seabed aspect, slope, and curvature, as well as several measures of terrain variability retained to encapsulate independent facets of topographic complexity, and calculated on both small (100 m) and large (1000 m) scales (Lecours et al., 2017; Supplementary Table S3). The topographic position index (TPI) was preferred over the relative difference to mean value (Lecours et al., 2017) as it could be more easily calculated at different spatial scales, given available software. Sea surface temperature, salinity, and fluorescence (a proxy for primary productivity) were also initially considered but not retained due to collinearity, inadequate resolution, insufficient spatial coverage, or a combination of the above. ...
The conservation of marine biodiversity is firmly embedded in national and international policy frameworks. However, the difficulties associated with conducting broad-scale surveys of oceanic environments restrict the evidence base available for applied management in pelagic waters. For example, the Oceanic Shoals Australian Marine Park (AMP) was established in 2012 in a part of Australia's continental shelf where unique topographic features are thought to support significant levels of biodiversity, yet where our understanding of ecological processes remains limited. We deployed mid-water baited remote underwater video systems (mid-water BRUVS) in the Oceanic Shoals AMP to provide the first non-extractive baseline assessment of pelagic wildlife communities in the area. We used these observations and high-resolution multibeam swaths of the seafloor to explore potential relationships between prominent geomorphological features and the (i) composition, (ii) richness, and (iii) relative abundance of pelagic communities. We documented 32 vertebrate species across three sampling areas, ranging from small baitfish to large sharks and rays, and estimated that up to nearly twice as many taxa may occur within the region as a whole. This highlights the Oceanic Shoals AMP as a reservoir of biodiversity comparable to other documented offshore oceanic hotspots. Our results also confirm the AMP as a possible distant foraging destination for IUCN red listed sea turtles, and a potential breeding and/or nursing ground for a number of charismatic cetaceans. Model outputs indicate that both species richness and abundance increase in proximity to raised geomorphic structures such as submerged banks and pinnacles, highlighting the influence of submarine topography on megafauna distribution. By providing a foundational understanding of spatial patterns in pelagic wildlife communities throughout a little studied region, our work demonstrates how a combination of non-destructive sampling techniques and predictive models can provide new opportunities to support decision-making under data shortage.
... The selection of environmental covariates is the first and key step in predictive mapping approaches based on the variable-environment relationship (Behrens et al. 2010;Lecours et al. 2017;Reichenbach et al. 2018). There exist a large number of potential environmental covariates. ...
... There exist a large number of potential environmental covariates. For example, at least 230 topographic covariates mentioned by Lecours et al. (2017) have been applied to predictive mapping of environment, and there are at least 596 covariates used for landslide susceptibility models alone in previous studies (Reichenbach et al. 2018). Clearly, what and how many covariates should be selected under different conditions is a major challenge to the application of these predictive mapping approaches, particularly for the non-experts. ...
Predictive mapping of environment is an important means for environment assessment and management. The selection of predictor variables (or environmental covariates) is the first and key step in predictive mapping. A number of machine learning and statistical models have been developed to select what and how many environmental covariates in a wide range of predictive mapping. Nevertheless, those models require a large amount of field data for model training and calibration, which can be problematic in applying to the areas with no or very limited field data available. To overcome the shortcoming, this paper proposes the most similar case method for selecting environmental covariates for predictive mapping. First, we describe the basic idea and the development procedures of the most similar case method; second, as an experimental test, we employ the proposed method to select the topographic covariates for inputting to the predictive soil mapping; third, we evaluate the effectiveness of the proposed method in the designed experiment using the leave-one-out cross-validation method. In total, 191 evaluation cases are included in the experimental case base and the test results show that 58.7% of the topographic covariates originally used in each evaluation case are correctly selected by the proposed method, which suggests that the proposed most-similar-case method perform reasonably well even with a relatively limited size of the case base. The future work should include the selection of other types of environmental covariates (e.g., climate, organism, etc.) and the development of an automatic method to extract the existing application cases from literature.
... With the increasing availability of different GIS tools, more and more topographic attributes have been successfully used as DSM covariates (McBratney et al. 2003;Bishop and Minasny 2016). However, because of the complexity of digital terrain modeling (Hengl and Reuter 2008) and the lack of guidance on the condition of using each individual topographic attribute as DSM covariate, non-expert DSM users who do not understand the functionality of each of the large number of available topographic attributes may select an arbitrary and sub-optimal set of topographic covariates for their DSM applications (Lecours et al. 2017). Therefore, the current design of the proposed method is focused on the recommendation of topographic covariates, which illustrates the concept simply and without loss of generality. ...
... The current method only considers the complexity of terrain; for now, it does not include other factors which describe the study area, such as land use, geology, and climate. This is because the current design of the proposed method is focused on selecting topographic covariates for DSM, and the complexity of terrain is usually used for representing the topographic characteristics (e.g., plain or mountain) which could often influence the selection of topographic covariates (Lecours et al. 2017). In this study, a set of quantitative attributes including total relief, standard deviation of elevation, and mean slope ) are adopted to describe the complexity of terrain in the mapping area. ...
Selecting a proper set of covariates is one of the most important factors that influence the accuracy of digital soil mapping (DSM). The statistical or machine learning methods for selecting DSM covariates are not available for those situations with limited samples. To solve the problem, this paper proposed a case-based method which could formalize the covariate selection knowledge contained in practical DSM applications. The proposed method trained Random Forest (RF) classifiers with DSM cases extracted from the practical DSM applications and then used the trained classifiers to determine whether each one potential covariate should be used in a new DSM application. In this study, we took topographic covariates as examples of covariates and extracted 191 DSM cases from 56 peer-reviewed journal articles to evaluate the performance of the proposed case-based method by Leave-One-Out cross validation. Compared with a novices' commonly-used way of selecting DSM covariates, the proposed case-based method improved more than 30% accuracy according to three quantitative evaluation indices (i.e., recall, precision, and F1-score). The proposed method could be also applied to selecting the proper set of covariates for other similar geographical modeling domains, such as landslide susceptibility mapping, and species distribution modeling.
... However, details on the use of terrain attributes derived from digital terrain models (DTMs) using neighbourhood operations (e.g., slope, rugosity, orientation, curvature) in SDM has received less attention. Deriving terrain attributes has become a routine operation, despite several potential pitfalls in the data processing workflow (Lecours et al., 2017b). ...
... Ecologists and biogeographers are facing an increasingly overwhelming amount and diversity of data and tools, often with little guidance to inform methodological and technical decisions (e.g., Lecours et al., 2017b;Michener and Jones, 2012). In this paper, we illustrate that critical information related to scale alteration approach of DTM terrain attributes (e.g., slope) is often undocumented in ecological studies, and we emphasize the effects that two standard scale alteration approaches can have on an analytical output. ...
Terrain attributes (e.g., slope, rugosity) derived in Geographic Information Systems (GIS) from digital terrain
models (DTMs) are widely used in both terrestrial and marine ecological studies due to their potential to act as
surrogates of species distribution. However, the spatial resolution of DTMs is often altered to match the scale at
which species observations were collected. Here, we highlight the significance of adequately reporting the
methods used to derive terrain attributes from DTMs and the consequences of their incorrect reporting in
ecological studies. To ensure full repeatability of studies, they should report (i) the source and the resolution of
the original DTM; (ii) the algorithm used to calculate terrain attributes; (iii) the method used for rescaling (e.g.,
aggregating or resampling, using the mean or maximum values); and (iv) the order in which these operations
were performed. We contrast the effects of two common scale alteration approaches for the derivation of terrain
attributes from DTMs. These two scale alteration methods differ in the step at which the change is performed: (i)
the resolution alteration is performed after computing terrain attributes from the original DTM at the native
resolution, or (ii) the resolution alteration is performed on the native DTM before computing terrain attributes.
While these approaches conceptually do the same thing (i.e., change the resolution of the terrain attributes), we
demonstrate that they produce two distinct sets of variables that are not interchangeable and describe different
properties of the terrain. In a species distribution modelling (SDM) context, the first approach calculates terrain
attribute values within the cell where a species is found, while the second approach calculates terrain attribute
values with respect to neighbouring cells. A mutual substitution of the two approaches results in a decrease of
models' discrimination ability and in misleading spatial predictions of species probability of occurrence.
Regardless of the DTM-derived attribute, we argue that the choice of the approach should be carefully guided by
both the ecological scale relevant to the question being asked and the performance of pre-analyses. We emphasize
that selected methods be clearly described to encourage reproducibility and proper interpretation of
results, thus enabling a better understanding of the role of scale in ecology.
... The cleaned acoustic data for each frequency were also imported to QPS FMGT for backscatter processing, and five surfaces were produced, also at 1 m horizontal resolution. The five bathymetric datasets were used to generate 15 terrain attributes: slope, easterness and northerness, maximum, minimum and mean curvatures, planform and profile curvatures, twisting curvature [5], a topographic position index [6], the relative difference from mean value [7], a vector ruggedness measure [8], a surface area to planar area ratio [9], an adjusted standard deviation metric [10], and the roughness indexelevation [11]. In addition, the datasets were used to classify the study area into seven morphometric features [12]: channels, passes, peaks, pits, planar flat areas, planar slope areas, and ridges. ...
In recent years, new multibeam echosounders that can simultaneously collect data at multiple frequencies have become available. However, the effects of acoustic frequency on bathymetric data have yet to be characterized, as early research on these new systems has instead focused on backscatter data. Here we explore such effects by deriving terrain attributes and classifications from bathymetric data from Head Harbour, Nova Scotia, Canada, that were collected at five different operating frequencies. The geomorphometric analyses were conducted on bathymetric surfaces generated from data collected at each operating frequency using four scales of analysis. Results show that bathymetry, its derived terrain attributes, and terrain classifications produced with them are all dependent on the acoustic frequency used to collect bathymetric data. While the observed effects on the regional bathymetry were relatively minor, local bathymetry, terrain attributes and terrain classifications were highly impacted by the frequency used when collecting data. The impacts were less important when the terrain attributes and classifications were generated using broader scales of analysis. These results raise questions about how bathymetry is measured and defined and how we should interpret the outcomes of marine geomorphometric analyses. This is particularly relevant as such analyses have become a key component of marine habitat mapping and submarine geomorphology mapping.
... Protected areas have increased globally under the call of the Convention of Biological Diversity (CBD) to improve the status of biodiversity by protecting at least ten percent of coastal and marine areas by 2020 (CBD, 2010). However, many protected areas are located in residual sites of low interest to fisheries and low biodiversity and ecological value (Lecours et al., 2017). Analogously, many others are not well enforced; for instance, establishing protected areas does not guarantee their effectiveness (Rife et al., 2013). ...
Anthropogenic impacts on the deep sea have increased in the last decades, representing a threat to vulnerable marine ecosystems (VMEs) and fisheries resources. Several measures to conserve VMEs at the local and international level were created, such as marine protected areas and fisheries moratoriums. VMEs are important habitats and are suggested to offer refuge to commercial species from fisheries. Hence VMEs conservation may benefit fisheries resources. This literature review identifies and assesses methods and evidence used to determine the performance of VME conservation measures at protecting fisheries resources. Seven studies conducted in the Arctic, Northeast and Central Atlantic, Mediterranean Sea, Central and West Pacific were identified. These studies used various survey methods to provide relevant information regarding the effectiveness of VME conservation measures. Survey methods include fishing gear, multibeam echosounder, side-scan sonar, tagging and recapture, telemetry, underwater video, and autonomous baited landers. The most used method in the literature reviewed was fishing gear. However, due to VME fragility, less invasive survey methods should be preferred. Evidence of protection includes changes in fish body size which seem to increase across the years of enforcement, and increased abundance, or density, of certain commercial species inside conserved VMEs.
... Expressed in indices that indicate geometric characteristics, the drainage network and the relief with it is strengthened by the analysis possibilities provided by geographic information systems (GIS) and remote sensing (RS) [29][30][31]. ...
Protected areas (PA) play an important role in minimizing the effects of soil erosion in watersheds. This study evaluated the performance of machine learning models, specifically support vector machine with linear kernel (SVMLinear), support vector machine with polynomial kernel (SVMPoly), and random forest (RF), on identifying indicators of soil erosion in 761 sub-watersheds and PA in northern Portugal, by using soil erosion by water in Europe, according to the revised universal soil loss equation (RUSLE2015), as target variable. The parameters analyzed were: soil erosion by water in Europe according to the revised universal soil loss equation (RUSLE2015), total burned area of the sub-watershed in the period of 1975-2020, fire recurrence, topographic wetness index (TWI), and the morphometric factors, namely area (A), perimeter (P), length (L), width (W), orientation (O), elongation ratio (Re), circularity ratio (Rc), compactness coefficient (Cc), form factor (Ff), shape factor (Sf), DEM, slope, and curvature. The median coefficient of determination (R 2) for each model was RF (0.61), SVMpoly (0.68), and SVMLinear (0.54). Regarding the analyzed parameters , those that registered the greatest importance were A, P, L, W, curvature, and burned area, indicating that an analysis which considers morphometric factors, together with soil erosion data affected by water and soil moisture, is an important indicator in the analysis of soil erosion in watersheds .
... Several products derived from RS (Remote Sensing) surveys are increasingly available for environmental studies and such information has shown to be extremely valuable for spatiotemporal assessment (e.g. Lecours et al., 2017;Murray et al., 2018;Pettorelli et al., 2018;Klein et al., 2021). That is due to several advantages of RS data, including easiness and velocity to obtain and share, free availability for part of them, standardization (which makes possible an easy way to develop and adjust computational tools for processing them), quality control, spatial coverage, availability for large areas (Smith & Clark, 2005). ...
Topographic data is increasingly available from LiDAR (Light Detection and Ranging) surveys. This research evaluates the limitations and capabilities of a LiDAR Digital Terrain Model (DTM) regarding catchment topography representation and river drainage network derivation, considering high-resolution (1 m) and resampled versions (2, 5, 10, and 30 m), and the Garças river basin (4,100 km²; Pernambuco state) as a study case. The terrain representation of the 1m-DTM and the derived network present outstanding quality, and its coarsening up to 30m resolution still outperforms the results obtained with SRTM data. LiDAR DTM coarsened to 2, 5, 10 and 30 m led to river length shortening of 0.1%, 0.3%, 1.2%, and 4%, respectively, while the difference between LiDAR 1m and SRTM was about 12%. The computational cost for 1m-DTM processing was prohibitive when using a typical low-cost computer, while some algorithms proved to be largely efficient (100 times faster) when running on a more powerful machine. DTM coarsening is an alternative to achieving a better balance between data quality and computer requirements.
Keywords:
Digital terrain model; Resampling; Flow directions; Digital elevation model; Computational cost
... In addition, we calculated the distance to highways, roads, trails, and rivers. The terrain variables (Lecours et al., 2017) were derived from a high-accuracy digital terrain model at a 25-m resolution (Moudrý et al., 2018). The terrain attributes were derived using Horn's algorithm at that resolution and subsequently aggregated to the analysis grain using the mean value from each cell (Moudrý et al., 2019). ...
Species distribution models (SDMs) are powerful tools in ecology and conservation. Choosing the right environmental drivers and filtering species' occurrences taking their biases into account are key factors to consider before modeling. In this case study, we address five common problems arising during the selection of input data for presence-only SDMs on an example of a general-ist species: the endangered Cantabrian brown bear. First, we focus on the selection of environmental variables that may drive its distribution, testing if climatic variables should be considered at a 1-km analysis grain. Second, we investigate how filtering the species' data in view of (1) their collection procedures , (2) different time frames, (3) dispersal areas, and (4) subpopulations affects the performance and outputs of the models at three different spatial analysis grains (500 m, 1 km, and 5 km). Our results show that models with different input data yielded only minor differences in performance and behaved properly in terms of model validation, although coarsening the analysis grain deteriorated model performance. Still, the contribution of individual variables and the habitat suitability predictions differed among models. We show that a combination of limited data availability and poor selection of environmental variables can lead to inaccurate predictions. Specifically for the brown bear, we conclude that climatic variables should not be considered for exploring habitat suitability and that the best input data for modeling habitat suitability in the study area originate from (1) observations and traces from the (2) most recent period (2006-2019) in which the population is expanding, (3) not considering cells of dispersing bear occurrences and (4) modeling sub-populations independently (as they show distinct habitat preferences). In conclusion , SDMs can serve as a useful tool for generalist species including all available data; still, expert evaluation from the perspective of data suitability for the purpose of modeling and possible biases is recommended. This is especially important when the results are intended for management and conservation purposes at the local level, and for species that respond to the
environment at coarse analysis grains.
... It is also based on the multichannel feature extraction, multichannel feature fusion, and landform recognition networks, as shown in Fig. 3. Relative difference to mean value is correlated with the topographic position index and several curvature types (general, planar, minimum, maximum, and average). It is a relative position index that identifies peaks (positive values) and craters (negative values) [44]. Input parameters are unmodified default values. ...
Deep learning networks have facilitated the automated scene recognition of landforms based on geomorphogenesis. However, current genetic landform classification methods do not consider regional geological context, which can more accurately reflect the formation and evolution mechanism of geomorphic landforms than local ones. Therefore, this study proposes a multimodal, deep-learning landform recognition framework based on a joint contextual geological and channel attention module (GCMENET). First, the multi-branch feature extraction network of DenseNet121 is used to extract the respective features from the target scene and the contextual geological scene. Second, the features similar to the landform features of the target scene are extracted from the contextual geological features based on the cosine method and then combined with geomorphic features of the target scene. Third, channel attention mechanism is used to reduce the interference caused by redundant contextual geological information after fusion of data. To measure the classification accuracy of GCMENET, we establish a fine geomorphogenic dataset consisting of remote sensing images of six landform types with a 64 × 64-pixel size and 10 m resolution (JOS10m). During the training process of two geomorphogenic datasets, the feature extraction network without batchnorm2d (batch normalization) could preserve the distribution and spatial alignment of data from the components. Using different training-to-validation data ratios and combinations of input components, the results of the GCMENET supplemented with the joint contextual geological and channel attention module exhibited greater accuracy than those obtained without the module. This observation confirms the importance of contextual geological information in automated geomorphogenic landforms.
... In the case of natural heterogeneity, one of the main challenges was to combine information from different spatial resolutions. In terms of land division, although there are different ways of defining cell size (Ciglič and Perko, 2017;Lecours et al., 2017), we did not compare different methods. It may be important to use several different methods of landscape analysis for adequate validation of the results. ...
Extensive agriculture imprints a certain spatial and temporal homogeneity on a landscape and often hides its natural heterogeneity. Moreover, this homogenization may be disconnected from land use capacity, increasing the risk of land degradation. We described the anthropogenic and natural heterogeneity of an agricultural basin of the central Rolling Pampa of Argentina, to assess whether the current land use is consistent with the natural land capacity. First, we mapped and characterized the anthropogenic and natural spatial heterogeneity through two independent analyses, one based on land cover and the other on geomorphological and edaphic variables. Second, we investigated the relationship between anthropogenic and natural heterogeneity by integrating the previous analyses with a third one, focused on vegetation activity (normalized difference vegetation activity index, NDVI). The results showed two anthropogenic landscape units, whereas six natural landscape units were identified. Anthropogenic landscape I was mainly associated with natural landscapes 1, 2 and 3. These landscapes were characterized by large, regularly shaped fields dominated by soybean. Anthropogenic landscape II was associated with natural landscapes 4, 5 and 6, which were characterized by smaller and more irregularly shaped fields and were more diversely cultivated than the other group of landscapes. Anthropogenic landscape I and its three associated natural landscapes had lower average vegetation activity but higher seasonal and inter-annual variation than anthropogenic landscape II and its associated landscapes. In conclusion, on a landscape level, natural heterogeneity may be hidden in agricultural landscapes with highly homogeneous land use, such as the Pampas region. If not taken into consideration, this could increase the risk of land degradation in certain conditions or landscapes.
... The segmentation implemented in RSOBIA requires the user to define both the number and the minimum size of the clusters, which are then classified by a k-means clustering approach. Another ArcGIS tool is Terrain Attribute Selection for Spatial Ecology [108], which derives six variables from bathymetry that can be used along with backscatter data in a multiresolution segmentation algorithm requiring expert judgment to define scale and input variables. Both the aforementioned approaches in ArcGIS can offer good predictive performance, but they require proprietary software, some degree of expert handling for fine tuning the model and, finally, they both suffer from the same problems as random forest when dealing with complex seafloors [30]. ...
Seafloor topography and grain size distribution are pivotal features in marine and coastal environments, able to influence benthic community structure and ecological processes at many spatial scales. Accordingly, there is a strong interest in multiple research disciplines to obtain seafloor geological and/or habitat maps. The aim of this study was to provide a novel, automatic and simple model to obtain high-resolution seafloor maps, using backscatter and bathymetric multibeam system data. For this purpose, we calibrated a linear regression model relating grain size distribution values, extracted from samples collected in a 16 km2 area near Bagnoli–Coroglio (southern Italy), against backscatter and depth-derived covariates. The linear model achieved excellent goodness-of-fit and predictive accuracy, yielding detailed, spatially explicit predictions of grain size. We also showed that a ground-truth sample size as large as 40% of that considered in this study was sufficient to calibrate analogous regression models in different areas. Regardless of some limitations (i.e., inability to predict rocky outcrops and/or seagrass meadows), our modeling approach proved to be a flexible tool whose main advantage is the rendering of a continuous map for sediment size, in lieu of categorical mapping approaches which usually report sharp boundaries or rely on a few sediment classes.
... The geomorphology of the seabed can provide important clues as to the dominant (local) physical processes, such as whether current and hydrodynamic regimes will enhance erosion or deposition. This information is commonly used within an ecological context, for instance, habitat modelling studies (Wilson et al., 2007;Lecours et al., 2017) and recently, has been used to predict the distribution of seabed sediments (Misiuk et al., 2018). As already noted, OC is associated with fine-grained material (Hedges and Keil, 1995), which is more likely to settle in low-energy environments; thus, it may be possible to identify relationships between seabed terrain attributes or geomorphological features that may indicate a depositional environment. ...
Marine sediments hold vast stores of organic carbon (OC). Techniques to spatially map sedimentary OC must develop to form the basis of seabed management tools that consider carbon-rich sediments. While the natural burial of carbon (C) provides a climate regulation service, the disturbance of buried C could present a significant positive feedback mechanism to atmospheric greenhouse gas concentrations. We present a regional Scottish case study that explores the suitability of integrating archived seafloor acoustic data (i.e., multibeam echosounder bathymetry and backscatter) with physical samples toward improved spatial mapping of surface OC in a dynamic coastal environment. Acoustic backscatter is a proxy for seabed sediments and can be collected over extensive areas at high resolutions. Sediment type is also an important predictor of OC. We test the potential of backscatter as a proxy for OC which may prove useful in the absence of exhaustive sediment data. Overall, although statistically significant, correlations between the variables OC, sediment type, and backscatter are relatively weak, likely reflecting a combination of limited and asynchronous data, sediment mobility over time, and complex environmental processing of OC in shelf sediments. We estimate linear mixed models to predict OC using backscatter and Folk sediment type as covariates. Our results show that incorporating backscatter in the model improves the precision of OC predictions by 14%. Backscatter discriminates between coarse and fine sediments, and therefore low and high OC regimes; however, was not able to discriminate amongst finer sediments. Although sediment type is a stronger predictor of OC, these data are available at a much lower spatial resolution and do not account for fine-scale variability. The resulting maps display varying spatial distributions of OC reflecting the different scales of the predictor variables, demonstrating a need for further methodological development. Backscatter shows considerable promise as a high-resolution predictor variable to improve the precision of surface OC maps, or to reduce the number of OC measurements required to achieve a specified precision. Applications of such maps have potential in improved C-stock estimates and the design of conservation and management strategies that consider marine sediments as valuable C stores.
... В работах, использующих аппарат геоморфометрии, применяется расчет, визуализация и анализ цифровых моделей и карт морфометрических величин [Florinsky, 2016[Florinsky, , 2017a. Описано несколько десятков морфометрических величин, причем многие из них известны под разными названиями [Hengl, Reuter, 2009;Lecours et al., 2016bLecours et al., , 2017cFavalli, Fornaciai, 2017;Amatulli et al., 2018Amatulli et al., , 2020Newman et al., 2018;Wilson, 2018;Minár et al., 2020]. Для успешного применения методов геоморфометрии пользователь должен ориентироваться в этих величинах, понимать их физико-математический смысл и знать их интерпретации. ...
Topography is the most important component of the geographical shell, one of the main elements of geosystems, and the framework of a landscape. geomorphometry is a science, the subject of which is modeling and analyzing the topography and the relationships between topography and other components of geosystems. Currently, the apparatus of geomorphometry is widely used to solve various multi-scale problems of the Earth sciences. As part of the RFBR competition “Expansion”, we present an analytical review of the development of theory, methods, and applications of geomorphometry for the period of 2016–2021. For the analysis, we used a sample of 485 of the strongest and most original papers published in international journals belonging to the JCR Web of Science Core Collection quartile I and II (Q1–Q2), as well as monographs from leading international publishers. We analyze factors caused a progress in geomorphometry in recent years. These include widespread use of unmanned aerial survey and digital photogrammetry, development of tools and methods for survey of submarine topography, emergence of new publicly available digital elevation models (DEMs), development of new methods of DEM preprocessing for their filtering and noise suppression, development of methods of two-dimensional and three-dimensional visualization of DEMs, introduction of machine learning techniques, etc. We consider some aspects of the geomorphometric theory developed in 2016–2021. We discuss new computational methods for calculating morphometric models from DEM, as well as the problems facing the developers and users of such methods. We consider application of geomorphometry for solving multiscale problems of geomorphology, hydrology, soil science, geology, glaciology, speleology, plant science and forestry, zoogeography, oceanology, planetology, landslide studies, remote sensing, urban studies, and archaeology.
... Among these environmental covariates, the most widely used covariates are the terrain covariates. Because of the high availability of multi-source digital elevation model (DEM) data and the digital terrain analysis tools, there exist a large number of terrain covariates (Lecours, Devillers, Simms, Lucieer, & Brown, 2017;Wilson, 2018). For many DSM users (especially beginners), an initial challenge during DSM applications is how to properly select terrain covariates among many potential covariates. ...
The terrain covariate is one type of key environmental covariate used in digital soil mapping (DSM). Because the selection of proper terrain covariates relies largely on users’ DSM knowledge and application context, automatic selection of terrain covariates is valuable for DSM users (especially non-experts). Case-based reasoning provides a promising solution to this demand. Recently, two case-based reasoning strategies have been proposed (i.e., the covariate-classification strategy and the most-similar-case strategy). However, there is no comparison of their performance, as well as that of the implemented methods based on the same or different strategies. This study fills this knowledge gap through a comparison experiment on the two representative methods for each strategy, respectively. Experiments adopted a DSM case base including 191 cases, which involve 38 terrain covariates. Results from a cross-validation and a practical DSM application showed that the random forests method adopting the covariate-classification strategy performed best.
... In addition to bathymetric and backscatter data, secondary features (e.g. slope, roughness, backscatter entropy, etc.) are routinely used to predict seafloor characteristics (Colenutt et al., 2013;Bouchet et al., 2015;Lecours et al., 2017;Janowski et al., 2020). For example, in areas with rocky seafloor, rock platforms can be identifiable from the bathymetry and slope (Colenutt et al., 2013), and higher backscatter intensity (Feldens et al., 2019), and could represent a proxy for relatively higher bed shear stress (Siwabessy et al., 2018). ...
Seafloor substrate mapping has become increasingly important to guide the management of marine ecosystems. Full coverage substrate maps, however, cannot easily be created from point samples (e.g. grabs, videos) as a result of the time required for collection and their discrete spatial extent. Instead, relationships between substrate types and surrogate variables as obtained from bathymetric or backscatter data can be modelled to build predictive substrate maps. As calculation of these surrogate variables is scale-dependent, the scale(s) of analysis need(s) to be selected first, with multiple scales likely required to adequately capture substrate characteristics. This paper proposes an objective and automatic self-adaptive analysis scale determination approach at each bathymetric point to extract terrain features (e.g. slope, aspect, etc). Object-based image analysis (OBIA) is also used to calculate additional texture features for segmented backscatter image objects. Random Forest classification is then used to model the relationship between these extracted features and substrate types interpreted from ground-truth video data, and full-coverage seafloor substrate maps are produced. The proposed method was applied on two datasets from Newfoundland, Canada, and demonstrated good performance in terms of both overall (>80%) and per-class accuracies. The proposed method is easily transferable to other study areas and provides an objective, repeatable means for classifying seafloor substrates for environmental protection and management of marine habitats.
... The seabed morphology was characterized following Lecours et al. (2017), using the following terrain variables derived from the 500 m bathymetry raster: local mean depth, slope, aspect (divided into northness and eastness), bathymetric position index (BPI), and vector ruggedness. ...
... Six derivatives of bathymetry were created using the Terrain Attribute Selection for Spatial Ecology (TASSE) tool [50] based on a circular radius of 25 cells. Some noise was present in the bathymetry data, so this radius was selected to characterise the local depth variation rather than the noise. ...
The ocean floor, its species and habitats are under pressure from various human activities. Marine spatial planning and nature conservation aim to address these threats but require sufficiently detailed and accurate maps of the distribution of seabed substrates and habitats. Benthic habitat mapping has markedly evolved as a discipline over the last decade, but important challenges remain. To test the adequacy of current data products and classification approaches, we carried out a comparative study based on a common dataset of multibeam echosounder bathymetry and backscatter data, supplemented with groundtruth observations. The task was to predict the spatial distribution of five substrate classes (coarse sediments, mixed sediments, mud, sand, and rock) in a highly heterogeneous area of the southwestern continental shelf of the United Kingdom. Five different supervised classification methods were employed, and their accuracy estimated with a set of samples that were withheld. We found that all methods achieved overall accuracies of around 50%. Errors of commission and omission were acceptable for rocky substrates, but high for all sediment types. We predominantly attribute the low map accuracy regardless of mapping approach to inadequacies of the selected classification system, which is required to fit gradually changing substrate types into a rigid scheme, low discriminatory power of the available predictors, and high spatial complexity of the site relative to the positioning accuracy of the groundtruth equipment. Some of these issues might be alleviated by creating an ensemble map that aggregates the individual outputs into one map showing the modal substrate class and its associated confidence or by adopting a quantitative approach that models the spatial distribution of sediment fractions. We conclude that further incremental improvements to the collection, processing and analysis of remote sensing and sample data are required to improve map accuracy. To assess the progress in benthic habitat mapping we propose the creation of benchmark datasets.
... 1. DEM was progressively smoothed, i.e., aggregated by a factor of two, resulting in six different resolution DEM rasters along a geometric series between 0.2 and 6.4 m (DEM1: 0.2 m, DEM2: 0.4 m, DEM3: 0.8 m, DEM4: 1.6 m, DEM5: 3.2 m, DEM6: 6.4 m), and another scale was also calculated to meet the resolution of the measuring campaigns (10 m, DEM7) 2. Terrain analysis was performed on each of the DEMs, giving firstly a series of terrain attribute rasters with different raster cell sizes/resolutions 3. Terrain attributes were then disaggregated, or upscaled to the original resolution of DEM1, resulting in differently smoothed terrain features with the same raster cell sizes Six terrain attributes were calculated following the guidelines 35 for the best characterization of the surface with the least potential co-variance between the selected attributes and the ones which were found to be applicable for a range of terrain complexities: Details about the calculations can be found in SI. ...
Soil respiration of grasslands is spatio-temporally variable reflecting the changing biological activities of the soil. In our study we analysed how the long-term soil respiration activities of dry grasslands would perform in terms of resistance and resilience. We also investigated how terrain features are responsible for response stability. We conducted a 7-year-long spatial study in a Hungarian dry grassland, measuring soil respiration (Rs), soil temperature (Ts) and soil water content (SWC) along 15 measuring campaigns in 80 × 60 m grids and soil organic carbon content in 6 of the occasions. Two proxy variables were introduced to grasp the overall Rs activity, as well as its temporal stability: average rankRs, the temporal average Rs rank of a measuring position from the campaigns revealed the persistent spatial pattern of Rs, while rangeRs, the range of ranks of the positions from the campaigns described the amplitude of the Rs response in time, referring to the response stability in terms of resistance or resilience. We formulated a hypothetic concept of a two-state equilibrium to describe the performance of the long-term Rs activity: Rs activity with smaller rangeRs, that is both the lower elevation positions with larger rankRs (“state I”) and the higher elevation positions with smaller rankRs (“state II”) correspond to an equilibrium state with several terrain attributes being responsible for the equilibrium responses. Majority of the measuring positions was belonging to none of these equilibrium states. These positions showed higher rangeRs for medium rankRs, suggesting resilience (not resistance) as a major strategy for this ecosystem.
... In addition, we tested the set of generalizable LSVs against six terrain variables (named LSM_tasse), proposed by Lecours et al. [13] to be used in environmental studies, as the six variables capture more than 70% of the topographic structure of an area. These variables are: relative difference to mean elevation value, standard deviation of DEM, easterness, northerness, local mean and slope. ...
... Berrill and O'Hara 2016;Bispo et al. 2016). The number of potential geomorphometric variables that could prove useful in this type of terrain analysis is large (Hengl and Reuter 2009;Florinsky 2017) and growing (Lecours et al. 2017). ...
Digital elevation model (DEM) data were used with climate data to estimate productivity in 19 Eucalyptus plantations in Minas Gerais state, Brazil. Typically, plantation and individual stand growth and productivity estimates, such as Site Index (SI) and Mean Annual Increment (MAI), are based on field measures of height, tree diameter and age. Using a Random Forest modelling approach, SI and MAI were related to: i) DEM-based geomorphometric variables, and ii) WorldClim historical macro-climatic measures. Three operational SI classes (high, medium and low productivity) in 180 stands were mapped with an overall accuracy of 91.6%. Medium and high productivity sites were the most accurately classified. Low productivity sites had 76.5% producer’s accuracy and 92.9% user’s accuracy, and were the most extensive in the study area. Such sites are considered of high importance from a plantation management perspective since additional forestry operations are likely required to address low productivity and growth.
... All other environmental data sets were adjusted to match the same projection and resolution using bilinear interpolation. The seabed morphology was characterized following Lecours et al. (2017), using the following parameters derived from the 500 m bathymetry grid: local mean depth, slope, aspect (divided into northness and eastness), bathymetric position index (BPI), and vector roughness. The BPI indicates if a particular pixel forms part of a positive or negative feature of the surrounding terrain (Wilson et al., 2007). ...
... All other environmental data sets were adjusted to match the same projection and resolution using bilinear interpolation. The seabed morphology was characterized following Lecours et al. (2017), using the following parameters derived from the 500 m bathymetry grid: local mean depth, slope, aspect (divided into northness and eastness), bathymetric position index (BPI), and vector roughness. The BPI indicates if a particular pixel forms part of a positive or negative feature of the surrounding terrain (Wilson et al., 2007). ...
In the deep waters of the Nordic Seas and adjacent areas, several benthic habitats such as cold-water coral reefs, coral gardens, and deep-sea sponge aggregations have been classified as vulnerable marine ecosystems (VMEs), due to their uniqueness, limited spatial extent, physical fragility, and slow recovery rate. In the last decade observations carried out by habitat mapping programmes in Norway, Iceland, and more recently in the Faroe Islands have substantially increased knowledge on the distribution of VMEs in the Nordic Seas. Nevertheless, large areas have not been explored due to the cost and logistics of obtaining observations in the deep-sea. Species distribution models can be used to predict the distribution of VMEs and their indicator species. Here we present the predicted distribution of 44 VME indicator taxa including 20 sponges, 17 cold-water corals, and 7 seapens in the Nordic Seas based on data compiled and models developed by the NovasArc project (2016-2018). Models for 44 VME indicator species were developed using the maximum entropy algorithm MaxEnt, using an extensive database compiled from habitat mapping surveys, by-catch data from bottom fish surveys, and records from reports and peer reviewed publications. Modeled distributions showed good agreement with observations. Niche overlap measures were used to identify seven groups and four subgroups of VME indicator taxa that co-occur. These were consistent with the species composition of known biotopes in the study area. A VME Index that combine the predictions for all VME indicator species was computed to identify particularly valuable and vulnerable ecosystems that should be targets of further exploration and conservation efforts. Such areas were identified at shelf break and slope off Iceland, the Faroe Islands, and central Norway, and the continental shelf off southern Greenland. The predicted distribution of VMEs in Arctic and sub-Arctic waters allows for the evaluation of interactions with fisheries and other anthropogenic activities and provides an important input for managers.
... While The Gully generally presented some minor differences (e.g., lower oxygen and nitrogen concentrations) when compared to the other two areas, those differences were not statistically significant. Following the approach detailed in Lecours et al. (2017), correlation and principal component analyses were then performed to reduce the number of variables used for the habitat mapping exercise. For example, those analyses identified relationships between temperature and nitrogen saturation and between slope and rugosity in this area. ...
This case study focuses on the characterization of fine-scale habitats associated with cold-water corals in three areas off Eastern Canada. Remotely operated vehicle (ROV)-based video, oceanographic, and bathymetric data were collected in 13 dives ranging from 200 to 3000 m deep at The Gully, the Flemish Cap, and the Orphan Knoll. Maps of potential habitats were produced and spatially compared with different taxa distributions, and species distribution models were computed to quantify the association of different environmental factors with cold-water corals. Results highlighted significant differences in corals' environmental preferences at all taxonomic levels. Results also showed the importance of collecting high-resolution chemical and oceanographic data as their integration with geomorphometric variables (e.g., aspect, rugosity, slope, topographic position) provides a more comprehensive picture of environmental niches. The exploration of cold-water coral habitats helped identify many cases of co-occurrences with a variety of other taxa including many deep-sea sponges.
... Geographic Information System (GIS) software usually offers tools to perform this task, unfortunately often using different algorithms for the same metric. Available software include GRASS, SAGA, Spatial Analyst toolset of ArcGIS, Benthic Terrain Modeller (Wright et al. 2012), ArcGeomorphometry (Rigol-Sanchez et al. 2015), BSG seabed mapping toolbox (De Clippele et al. 2016), TASSE (Terrain Attribute Selection for Spatial Ecology: Lecours et al. 2015Lecours et al. , 2017 and other software (Wood 2009). ...
The marine scientific community developed new habitat mapping procedures that proved to be effective to map seafloor habitats in a holistic manner, thus refining our knowledge of benthic sea life. Several works on this topic focused on cold-water coral habitats at the global scale, with focus on the Atlantic Ocean while, only a few studies applied these methods to the Mediterranean Sea. The prime objective of this chapter is to comment the methodologies used to describe cold-water coral habitats and map their extension; secondly we discuss statistical models to outline the most suitable areas for cold-water coral settlement (habitat suitability models). Furthermore, recent technological developments in habitat mapping procedures provide geophysical data and seafloor images of higher resolution than ever before, while multi-frequency systems supply new hyperspectral images of the seabed. This new type of data could be crucial for a better characterisation of such Mediterranean cnidarian habitats. Here we reviewed the procedures for the Mediterranean cold-water coral habitat mapping applied so far, describing two case studies in the Southern Adriatic Sea and discussing caveats and future perspectives.
... For this reason, we present a new application of BTM which substitutes ruggedness for depth in the BTM classification dictionary as it was determined to be the most relevant metric for defining outcropping features. These variables of TPI, slope, and ruggedness have been identified among the most effective variables in capturing topographic structure [68]. Ruggedness was determined to effectively capture outcropping features, which are a prominent feature in the case study area and NSW shelf seascapes more broadly. ...
In 2017, the New South Wales (NSW) Office of Environment and Heritage (OEH) initiated
a state-wide mapping program, SeaBed NSW, which systematically acquires high-resolution (2–5 m cell size) multibeam echosounder (MBES) and marine LiDAR data along more than 2000 km of the subtropical-to-temperate southeast Australian continental shelf. This program considerably expands upon existing efforts by OEH to date, which have mapped approximately 15% of NSW waters with these technologies. The delivery of high volumes of new data, together with the vast repository of existing data, highlights the need for a standardised, automated approach to classify seabed data. Here we present a methodological approach with new procedures to semi-automate the classification of high-resolution bathymetry and intensity (backscatter and reflectivity) data into a suite of data products including classifications of seabed morphology (landforms) and composition (substrates, habitats, geomorphology). These methodologies are applied to two case study areas representing newer (Wollongong, NSW) and older (South Solitary Islands, NSW) MBES datasets to assess the transferability of classification techniques across input data of varied quality. The suite of seabed classifications produced by this study provide fundamental baseline data on seabed shape, complexity, and composition which will inform regional risk assessments and provide insights into biodiversity and geodiversity.
... There is a long debate about which LSVs express better the surface characteristics. Many LSVs show collinearity [8] and Evans [9] states that the first and second derivatives of a DEM are enough in most cases. Extending the current approach to a multi-scale geomorphometric characterization as that proposed by Wood [10] is currently under investigation. ...
Attempts towards a global geomorphometric atlas have been done in the past when computational power was less evolved than nowadays. In this study, we present a possible way to create a global geomorphometric atlas by taking advantage of the Google Earth Engine (GEE) computational capabilities. To exemplify how accessible, efficient and fast GEE works in geomorphometric analysis we globally derive popular land-surface variables (e.g. slope, aspect, multiple curvatures, and surface roughness) and perform a hierarchical pixel-based classification of topography on three levels of details. The scale of analysis can easily be adjusted to meet the user needs. Performing large-scale morphometric analysis with GEE has the potential to become an important milestone in Geomorphometry and can lead towards generating a global geomorphometric atlas.
... Lecours et al. [54] demonstrated that six terrain attributes capture the majority of seafloor topographic structure: relative position, local standard deviation (as a measure of rugosity), slope orientation (easterness and northerness), local bathymetric mean, and slope. This combination of terrain attributes outperforms alternative combinations when used as abiotic surrogates for predictive marine habitat mapping [55]. ...
Seafloor mapping can offer important insights for marine management, spatial planning, and research in marine geology, ecology, and oceanography. Here, we present a method for generating regional bathymetry and geomorphometry maps from crowd-sourced depth soundings (Olex AS) for a small fraction of the cost of multibeam data collection over the same area. Empirical Bayesian Kriging was used to generate a continuous bathymetric surface from incomplete and, in some areas, sparse Olex coverage on the Newfoundland and Labrador shelves of eastern Canada. The result is a 75m bathymetric grid that provides over 100x finer spatial resolution than previously available for the majority of the 672,900 km² study area. The interpolated bathymetry was tested for accuracy against independent depth data provided by Fisheries and Oceans Canada (Spearman correlation = 0.99, p<0.001). Quantitative terrain attributes were generated to better understand seascape characteristics at multiple spatial scales, including slope, rugosity, aspect, and bathymetric position index. Landform classification was carried out using the geomorphons algorithm and a novel method for the identification of previously unmapped tributary canyons at the continental shelf edge are also presented to illustrate some of many potential benefits of crowd-sourced regional seafloor mapping.
Coastal and marine classifications, both spatially explicit in the form of maps and non-spatial representations of the environment, are critical to the effective implementation of ecosystem-based management strategies such as marine spatial planning. This chapter provides an overview of a wide range of coastal classifications and classified maps developed to simplify and communicate biological, physical, social, and economic patterns in support of enhanced management decision making. Examples are provided from around the world and span a range of spatial scales from global coastal classifications to those for individual bays and estuaries. Technological advances in remote sensing, social media analyses and artificial intelligence classifiers have diversified and improved the performance and application of biophysical coastal classifications and thematic maps. In the past decade, important progress has been made in the spatial representation of cultural values through participatory mapping and the integration of social- ecological patterns for coastal risk assessment. Mapping ecosystem services has become more widespread and the desire for environmental sustainability across sectors has seen a proliferation in predictive mapping applied to conservation prioritization and marine spatial planning. The chapter offers a showcase rather than a critique of applications and concludes with a section highlighting progress and future challenges in developing more culturally meaningful coastal classifications to inform coastal management.
Microevolutionary processes shape adaptive responses to heterogeneous environments, where these effects vary both among and within species. However, the degree to which signatures of adaptation to environmental drivers can be detected based on spatial scale and genomic marker remains largely unknown. We studied signatures of local adaptation across different spatial extents, investigating complementary types of genomic variants–single nucleotide polymorphisms (SNPs) and polymorphic transposable elements (TEs)–in populations of the alpine model plant species Arabis alpina. We coupled high-resolution (0.5m) environmental factors, derived from remote sensing digital elevation models, with whole-genome sequenced data of 304 individuals across four populations. We demonstrate that responses of A. alpina to similar amounts of abiotic variation are largely governed by local evolutionary processes and find minimally overlapping signatures of local adaptation between SNPs and polymorphic TEs. Notably, functional annotations of high-impact genomic variants revealed several defence-related genes associated with the abiotic factors studied, which could indicate indirect selective pressure of biotic agents. Our results highlight the importance of considering different spatial extents and types of genomic polymorphisms when searching for signatures of adaptation to environmental variation. Such insights provide key information on microevolutionary processes and could guide management decisions to mitigate negative impacts of climate change on alpine plant populations.
Pacific sand lance (Ammodytes personatus) support marine food webs in the Salish Sea, yet our knowledge of intertidal spawning habitat for this species is limited. Increasing participation in community science surveys for intertidal sand lance spawning has resulted in the detection of eggs on >90 beaches in the Canadian Salish Sea since 2001. Using this data, we developed a MaxEnt habitat suitability model using 6 environmental variables from a suite of 9. We estimate that only 5.4% of the intertidal zone of the Canadian Salish Sea has a moderate to high likelihood of providing suitable sand lance spawning habitat. This rare habitat was best predicted by its proximity to estuaries, shoreline slope, distance to predicted subtidal sand lance burying habitat, seabed substrate and aspect. Our model could be used as the basis for a Pacific coast-wide model in areas with less available information. Identifying intertidal spawning habitat of sand lance will support conservation efforts intended to maintain forage fish species.
Land-surface parameters derived from digital land surface models (DLSMs) (for example, slope, surface curvature, topographic position, topographic roughness, aspect, heat load index, and topographic moisture index) can serve as key predictor variables in a wide variety of mapping and modeling tasks relating to geomorphic processes, landform delineation, ecological and habitat characterization, and geohazard, soil, wetland, and general thematic mapping and modeling. However, selecting features from the large number of potential derivatives that may be predictive for a specific feature or process can be complicated, and existing literature may offer contradictory or incomplete guidance. The availability of multiple data sources and the need to define moving window shapes, sizes, and cell weightings further complicate selecting and optimizing the feature space. This review focuses on the calculation and use of DLSM parameters for empirical spatial predictive modeling applications, which rely on training data and explanatory variables to make predictions of landscape features and processes over a defined geographic extent. The target audience for this review is researchers and analysts undertaking predictive modeling tasks that make use of the most widely used terrain variables.
To outline best practices and highlight future research needs, we review a range of land-surface parameters relating to steepness, local relief, rugosity, slope orientation, solar insolation, and moisture and characterize their relationship to geomorphic processes. We then discuss important considerations when selecting such parameters for predictive mapping and modeling tasks to assist analysts in answering two critical questions: What landscape conditions or processes does a given measure characterize? How might a particular metric relate to the phenomenon or features being mapped, modeled, or studied? We recommend the use of landscape- and problem-specific pilot studies to answer, to the extent possible, these questions for potential features of interest in a mapping or modeling task. We describe existing techniques to reduce the size of the feature space using feature selection and feature reduction methods, assess the importance or contribution of specific metrics, and parameterize moving windows or characterize the landscape at varying scales using alternative methods while highlighting strengths, drawbacks, and knowledge gaps for specific techniques. Recent developments, such as explainable machine learning and convolutional neural network (CNN)-based deep learning, may guide and/or minimize the need for feature space engineering and ease the use of DLSMs in predictive modeling tasks.
Integrating Multibeam Echosounder (MBES) data (bathymetry and backscatter) and underwater video technology allows scientists to study marine habitats. However, use of such data in modeling suitable seagrass habitats in Malaysian coastal waters is still limited. This study tested multiple spatial resolutions (1 and 50 m) and analysis window sizes (3 × 3, 9 × 9, and 21 × 21 cells) probably suitable for seagrass-habitat relationships in Redang Marine Park, Terengganu, Malaysia. A maximum entropy algorithm was applied, using 12 bathymetric and backscatter predictors to develop a total of 6 seagrass habitat suitability models. The results indicated that both fine and coarse spatial resolution datasets could produce models with high accuracy (>90%). However, the models derived from the coarser resolution dataset displayed inconsistent habitat suitability maps for different analysis window sizes. In contrast, habitat models derived from the fine resolution dataset exhibited similar habitat distribution patterns for three different analysis window sizes. Bathymetry was found to be the most influential predictor in all the models. The backscatter predictors, such as angular range analysis inversion parameters (characterization and grain size), gray-level co-occurrence texture predictors, and backscatter intensity levels, were more important for coarse resolution models. Areas of highest habitat suitability for seagrass were predicted to be in shallower (
Dense aggregations of horse mussels exist in the Bay of Fundy and are thought to be associated with high biodiversity compared to surrounding habitats. Previous research show correlations between these aggregations and long narrow flow-parallel bedforms. In this study, high resolution seafloor photographs were compared to multibeam echosounder data to investigate benthic community and biodiversity patterns across the Bay of Fundy. Biophysical seafloor types (benthoscapes) were mapped with 71% accuracy using Object Based Image Analysis (OBIA) with fuzzy c-means classification. Community patterns were found to shift gradually across benthoscapes and mega-epifaunal biodiversity was highest in areas of mixed sediments and silty gravel and anemones. Horse mussels occurred mainly within the sand benthoscape where the flow parallel bedforms exist. Community patterns were also assessed at fine scales (5 m) using a by-eye interpretation of the flow parallel bedforms, where gradational shifts across the bedforms and surrounding areas were evident and abundance of horse mussels were found to be significantly greater on the bedforms. These findings will ultimately facilitate decisions around fisheries management, marine spatial planning, and monitoring of the horse mussel habitats and the surrounding benthoscapes within the Bay of Fundy.
Seafloor morphology at Mid-Ocean Ridges is thought to be controlled by the interaction of tectonic and magmatic processes. In turn, these are thought to control the distribution of volcanic and brittle tectonic features on the seafloor over multiple scales; from the placement of individual faults measuring hundreds of metres, to shallow and deep bathymetric swales over tens of kilometres. It is unclear, however, whether there is a systematic relationship between these smaller and larger features, with varying melt supply, spreading rate, and spreading obliquity. Here, to investigate this relationship systematically, I have developed an innovative method to robustly and objectively identify large scale variations in brittle tectonic extension, expressed on the seafloor as faults. This method allows comparison with a number of proxies for melt supply along-strike and orthogonally-away from spreading ridge segments. I objectively identify fault planes by cross-correlating dip, dip direction, and backscatter that might individually suggest their presence. The fault heave is measured to estimate extension over length and temporal scales comparable to those melt-related processes involved in forming second order ridge segments. Regions of high and low melt supply are identified and defined quantitatively from positive and negative anomalies in bathymetry. These indications are cross-validated by comparison to gravity anomalies and other bathymetric features diagnostic of high or low melt supply.
Following analysis of several mid-ocean ridge examples, spanning a range of spreading rates and ridge/spreading orientation, no consistent correlation between large scale extension and melt supply at scales of 1 km, 10 km, or 20 – 50 km was found. This null-result has three potential, contributing, explanations: First, environmental processes erode and mask faults, reducing their apparent heave which then deviates from the theoretical model. Second, fault extension occurs over a broader area surrounding the spreading axis while melt intrusion and delivery into the upper crust remains more focused on a narrow zone within the spreading axis. This would result in a spatial and temporal disconnect between the location and occurrence of brittle faulting and the location and timing of delivery of magma. Third, brittle extension and fault distribution may be locally controlled by dykes and lava domes, whose distribution may not scale-up to the larger magnitude of second order spreading centre features such as elevated and inflated segment centres and depressed and subsided segment ends.
Toolbox for ArcGIS built from results of Lecours et al. (2017) and validated in Lecours et al. (2016).
This toolbox generates six independent terrain attributes that together summarize topographic or bathymetric variability.
References:
Lecours, V., Devillers, R., Simms, A.E., Lucieer, V.L., and Brown, C.J. (2017) Towards a framework for terrain attribute selection in environmental studies. Environmental Modelling & Software, 89:19-30.
Lecours, V., Brown, C.J., Devillers, R., Lucieer, V.L., and Edinger, E.N. (2016) Comparing selections of environmental variables for ecological studies: a focus on terrain attributes. PLoS ONE, 11:e0167128.
TO CITE:
v. 1.1: Lecours, V. (2017) Terrain Attribute Selection for Spatial Ecology (TASSE)
v. 1.0: Lecours, V. (2015) Terrain Attribute Selection for Spatial Ecology (TASSE
Selecting appropriate environmental variables is a key step in ecology. Terrain attributes (e.g. slope, rugosity) are routinely used as abiotic surrogates of species distribution and to produce habitat maps that can be used in decision-making for conservation or management. Selecting appropriate terrain attributes for ecological studies may be a challenging process that can lead users to select a subjective, potentially sub-optimal combination of attributes for their applications. The objective of this paper is to assess the impacts of subjectively selecting terrain attributes for ecological applications by comparing the performance of different combinations of terrain attributes in the production of habitat maps and species distribution models. Seven different selections of terrain attributes, alone or in combination with other environmental variables, were used to map benthic habitats of German Bank (off Nova Scotia, Canada). 29 maps of potential habitats based on unsupervised classifications of biophysical characteristics of German Bank were produced, and 29 species distribution models of sea scallops were generated using MaxEnt. The performances of the 58 maps were quantified and compared to evaluate the effectiveness of the various combinations of environmental variables. One of the combinations of terrain attributes–recommended in a related study and that includes a measure of relative position, slope, two measures of orientation , topographic mean and a measure of rugosity–yielded better results than the other selections for both methodologies, confirming that they together best describe terrain properties. Important differences in performance (up to 47% in accuracy measurement) and spatial outputs (up to 58% in spatial distribution of habitats) highlighted the importance of carefully selecting variables for ecological applications. This paper demonstrates that making a subjective choice of variables may reduce map accuracy and produce maps that do not adequately represent habitats and species distributions, thus having important implications when these maps are used for decision-making.
Geomorphometry, the science of quantitative terrain characterization, has traditionally focused on the investigation of terrestrial landscapes. However, the dramatic increase in the availability of digital bathymetric data and the increasing ease by which geomorphometry can be investigated using geographic information systems (GISs) and spatial analysis software has prompted interest in employing geomorphometric techniques to investigate the marine environment. Over the last decade or so, a multitude of geomorphometric techniques (e.g. terrain attributes, feature extraction, automated classification) have been applied to characterize seabed terrain from the coastal zone to the deep sea. Geomorphometric techniques are, however, not as varied, nor as extensively applied, in marine as they are in terrestrial environments. This is at least partly due to difficulties associated with capturing, classifying, and validating terrain characteristics underwater. There is, nevertheless, much common ground between terrestrial and marine geomorphometry applications and it is important that, in developing marine geomorphometry, we learn from experiences in terrestrial studies. However, not all terrestrial solutions can be adopted by marine geomorphometric studies since the dynamic, four-dimensional (4-D) nature of the marine environment causes its own issues throughout the geomorphometry workflow. For instance, issues with underwater positioning, variations in sound velocity in the water column affecting acoustic-based mapping, and our inability to directly observe and measure depth and morphological features on the seafloor are all issues specific to the application of geomorphometry in the marine environment. Such issues fuel the need for a dedicated scientific effort in marine geomorphometry. This review aims to highlight the relatively recent growth of marine geomorphometry as a distinct discipline, and offers the first comprehensive overview of marine geomorphometry to date. We address all the five main steps of geomorphometry, from data collection to the application of terrain attributes and features. We focus on how these steps are relevant to marine geomorphometry and also highlight differences and similarities from terrestrial geomorphometry. We conclude with recommendations and reflections on the future of marine geomorphometry. To ensure that geomorphometry is used and developed to its full potential, there is a need to increase awareness of (1) marine geomorphometry amongst scientists already engaged in terrestrial geomorphometry, and of (2) geomorphometry as a science amongst marine scientists with a wide range of backgrounds and experiences.
Surveying primary tropical forest over large regions is challenging. Indirect methods of relating terrain information or other external spatial datasets to forest biophysical parameters can provide forest structural maps at large scales but the inherent uncertainties need to be evaluated fully. The goal of the present study was to evaluate relief characteristics, measured through geomorphometric variables, as predictors of forest structural characteristics such as average tree basal area (BA) and height (H) and average percentage canopy openness (CO). Our hypothesis is that geomorphometric variables are good predictors of the structure of primary tropical forest, even in areas, with low altitude variation. The study was performed at the Tapajós National Forest, located in the Western State of Pará, Brazil. Forty-three plots were sampled. Predictive models for BA, H and CO were parameterized based on geomorphometric variables using multiple linear regression. Validation of the models with nine independent sample plots revealed a Root Mean Square Error (RMSE) of 3.73 m2/ha (20%) for BA, 1.70 m (12%) for H, and 1.78% (21%) for CO. The coefficient of determination between observed and predicted values were r2 = 0.32 for CO, r2 = 0.26 for H and r2 = 0.52 for BA. The models obtained were able to adequately estimate BA and CO. In summary, it can be concluded that relief variables are good predictors of vegetation structure and enable the creation of forest structure maps in primary tropical rainforest with an acceptable uncertainty.
The performance of five methods for determining the number of components to retain (Horn's parallel analysis, Velicer's minimum average partial [MAP], Cattell's scree test, Bartlett's chi-square test, and Kaiser's eigenvalue greater than 1.0 rule) was investigated across seven systematically varied conditions (sample size, number of variables, number of components, component saturation, equal or unequal numbers of variables per component, and the presence or absence of unique and complex variables). We generated five sample correlation matrices at each of two sample sizes from the 48 known population correlation matrices representing six levels of component pattern complexity. The performance of the parallel analysis and MAP methods was generally the best across all situations. The scree test was generally accurate but variable. Bartlett's chi-square test was less accurate and more variable than the scree test. Kaiser's method tended to severely overestimate the number of components. We discuss recommendations concerning the conditions under which each of the methods are accurate, along with the most effective and useful methods combinations.
Neutral landscape models (NLMs) were developed from percolation theory nearly a decade ago. Since then, the original random percolation maps have undergone adaptive radiation and NLMs now include a diverse array of spatially explicit models based on theoretical distributions derived from fractal geometry and spectral synthesis. The purpose of NLMs is to provide null models of landscape structure as a baseline for comparison with real landscape patterns, or for evaluating the effects of landscape structure on ecological processes. As the use of NLMs has expanded beyond the domain of theoretical landscape ecology to applications in other areas of ecology, there is an increased risk that NLMs will be used inappropriately, or that their function will be misunderstood or misinterpreted. NLMs are being subjected to the same general criticisms levied against null models in other areas of ecology. For this reason, we clarify the purpose of NLMs, review the contributions of NLMs to ecology, and evaluate the appropriate use of NLMs in ecological research. NLMs have already made several contributions to ecology: (1) development of spatial indices to describe landscape patterns; (2) prediction of critical thresholds in ecological phenomena; (3) definition of landscape connectivity; (4) development of ''species' perceptions'' of landscape structure; (5) provision of a general model of spatial complexity; and (6) determination of the ecological consequences of spatial heterogeneity. In the future, emphasis on NLMs should shift from theoretical development to application and model testing.
Despite being identified as a driver of mobile predator aggregations (hotspots) in both marine and terrestrial environments, topographic complexity has long remained a challenging concept for scientists to visualise and a difficult parameter to estimate. It is only with the advent of high-speed computers and the recent popularisation of geographical information systems (GIS) that terrain attributes have begun to be quantitatively measured in three-dimensional space and related to wildlife dynamics, making the well-established field of geomorphometry (or ‘digital terrain modelling’) a discipline of growing appeal to biologists. Although a diverse array of numerical metrics is now available to describe the shape, geometry and physical properties of natural habitats, few of these are known to, or adequately used by, ecologists. In this review, we examine the nature and usage of 56 geomorphometrics extracted from the ecological modelling literature over a period of 32 years (1979–2011). We show that, in studies of mobile predators, numerous topographic variables have largely been overlooked in favour of single basic metrics that do not, on their own, fully capture the complexity of continuous landscapes. Based on a simulation approach, we assess the redundancy and correlation structure of these metrics and demonstrate that a majority are highly collinear. We highlight a suite of 7–8 complementary metrics which best explain topographic patterns across a bathymetric grid of the west Australian seafloor, and contend that field and analytical protocols should prioritise variables of these types, particularly when the responses of predator populations to physical habitat features are of interest. We suggest that prominent structures such as canyons, seamounts or mountain chains can serve as useful proxies for predator hotspots, especially in remote locations where access to high-resolution biological data is often limited.
In 2006 the United Nations called on states to implement measures to prevent significant adverse impacts to vulnerable marine ecosystems (VMEs) in the deep sea. It has been widely recognised that a major limitation to the effective application of these measures to date is uncertainty regarding the distribution of VMEs. Conservationists, researchers, resource managers and governmental bodies are increasingly turning to predictive species distribution models (SDMs) to identify the potential presence of species in areas that have not been sampled. In particular, the development of robust ‘presence-background’ model algorithms has accelerated the application of these techniques for working with presence-only species data. This has allowed scientists to exploit the large amounts of species data held in global biogeographic databases. Despite improvements in model algorithms, environmental data and species presences, there are still limitations to the reliability of these techniques, especially in poorly studied areas such as the deep sea. Recent studies have begun to address a key limitation, the quality of data, by using multibeam echosounder surveys and species data from video surveys to acquire high-resolution data. Whilst these data are often amongst the very best that can be acquired, the surveys are highly localised, often targeted towards known VME-containing areas, are very expensive and time consuming. It is financially prohibitive to survey whole regions or ocean basins using these techniques, so alternative cost-effective approaches are required. Here, we review ‘presence-background’ SDMs in the context of those studies conducted in the deep sea. The issues of sampling bias, spatial autocorrelation, spatial scale, model evaluation and validation are considered in detail, and reference is made to recent developments in species distribution modelling literature. Further information is provided on how these approaches are being used to influence ocean management, and best practises are outlined to aid the effective adoption of these techniques in the future.
Coral reefs represent one of the most irregular substrates in the marine environment. This roughness or topographic complexity is an important structural characteristic of reef habitats that affects a number of ecological and environmental attributes, including species diversity and water circulation. Little is known about the range of topographic complexity exhibited within a reef or between different reef systems. The objective of this study was to quantify topographic complexity for a 5-km x 5-km reefscape along the northern Florida Keys reef tract, over spatial scales ranging from meters to hundreds of meters. The underlying dataset was a 1-m spatial resolution, digital elevation model constructed from lidar measurements. Topographic complexity was quantified using a fractal algorithm, which provided a multi-scale characterization of reef roughness. The computed fractal dimensions (D) are a measure of substrate irregularity and are bounded between values of 2 and 3. Spatial patterns in D were positively correlated with known reef zonation in the area. Landward regions of the study site contain relatively smooth (D approximate to 2.35) flat-topped patch reefs, which give way to rougher (D approximate to 2.5), deep, knoll-shaped patch reefs. The seaward boundary contains a mixture of substrate features, including discontinuous shelf-edge reefs, and exhibits a corresponding range of roughness values (2.28 <= D <= 2.61).
Coral reef ecosystems are topographically complex environments and this structural heterogeneity influences the distribution, abundance and behavior of marine organisms. Airborne hydrographic lidar (Light Detection and Ranging) provides high resolution digital bathymetry from which topographic complexity can be quantified at multiple spatial scales. To assess the utility of lidar data as a predictor of fish and coral diversity and abundance, seven different morphometrics were applied to a 4 m resolution bathymetry grid and then quantified at multiple spatial scales (i.e., 15, 25, 50, 100, 200 and 300 m radii) using a circular moving window analysis. Predictive models for nineteen fish metrics and two coral metrics were developed using the new statistical learning technique of stochastic gradient boosting applied to regression trees. Predictive models explained 72% of the variance in herbivore biomass, 68% of parrotfish biomass, 65% of coral species richness and 64% of fish species richness. Slope of the slope (a measure of the magnitude of slope change) at relatively local spatial scales (15-100 m radii) emerged as the single best predictor. Herbivorous fish responded to topographic complexity at spatial scales of 15 and 25 m radii, whereas broader spatial scales of between 25 and 300 m radii were relevant for piscivorous fish. This study demonstrates great utility for lidar-derived bathymetry in the future development of benthic habitat maps and faunal distribution maps to support ecosystem based management and marine spatial planning.
Given the proliferation of factor analysis applications in the literature, the present article examines the use of factor analysis in current published research across four psychological journals. Notwithstanding ease of analysis due to computers, the appropriate use of factor analysis requires a series of thoughtful researcher judgments. These judgments directly affect results and interpretations. The authors examine across studies (a) the decisions made while conducting exploratory factor analyses (N = 60) and (b) the information reported from the analyses. In doing so, they present a review of the current status of factor analytic practice, including comment on common errors in use and reporting. Recommendations are proffered for future practice as regards analytic decisions and reporting in empirical research.
Expectations for reporting factor analysis results as part of construct validation are explored in the context of emerging views of measurement validity. Desired practices are discussed regarding both exploratory factor analysis (e.g., principal components analysis) and confirmatory factor analysis (e.g., LISREL and EQS factor analyses). A short computer program for conducting parallel analysis is appended.
In the feature subset selection problem, a learning algorithm is faced with the problem of selecting a relevant subset of features upon which to focus its attention, while ignoring the rest. To a c hieve the best possible performance with a particular learning algorithm on a particular training set, a feature subset selection method should consider how the algorithm and the training set interact. We explore the relation between optimal feature subset selection and relevance. Our wrapper method searches for an optimal feature subset tailored to a particular algorithm and a domain. We study the strengths and weaknesses of the wrapper approach a n d s h o w a series of improved designs. We compare the wrapper approach to induction without feature subset selection and to Relief, a lter approach to feature subset selection. Signiicant improvement in accuracy is achieved for some datasets for the two families of induction algorithms used: decision trees and Naive-Bayes.
Background
Co-expression measures are often used to define networks among genes. Mutual information (MI) is often used as a generalized correlation measure. It is not clear how much MI adds beyond standard (robust) correlation measures or regression model based association measures. Further, it is important to assess what transformations of these and other co-expression measures lead to biologically meaningful modules (clusters of genes).
Results
We provide a comprehensive comparison between mutual information and several correlation measures in 8 empirical data sets and in simulations. We also study different approaches for transforming an adjacency matrix, e.g. using the topological overlap measure. Overall, we confirm close relationships between MI and correlation in all data sets which reflects the fact that most gene pairs satisfy linear or monotonic relationships. We discuss rare situations when the two measures disagree. We also compare correlation and MI based approaches when it comes to defining co-expression network modules. We show that a robust measure of correlation (the biweight midcorrelation transformed via the topological overlap transformation) leads to modules that are superior to MI based modules and maximal information coefficient (MIC) based modules in terms of gene ontology enrichment. We present a function that relates correlation to mutual information which can be used to approximate the mutual information from the corresponding correlation coefficient. We propose the use of polynomial or spline regression models as an alternative to MI for capturing non-linear relationships between quantitative variables.
Conclusion
The biweight midcorrelation outperforms MI in terms of elucidating gene pairwise relationships. Coupled with the topological overlap matrix transformation, it often leads to more significantly enriched co-expression modules. Spline and polynomial networks form attractive alternatives to MI in case of non-linear relationships. Our results indicate that MI networks can safely be replaced by correlation networks when it comes to measuring co-expression relationships in stationary data.
This chapter presents a broad synthesis and overview based on the 57 case studies. The case studies covered areas of seafloor ranging from 0.15 km2 to over 1,000,000 km2 and a broad range of geomorphic feature types. Of the many purposes for mapping benthic habitats, four stand out as preeminent: to support government spatial marine planning, management, and decision making; to support and underpin the design of marine protected areas (MPAs); to conduct scientific research programs aimed at generating knowledge of benthic ecosystems and seafloor geology; and to conduct living and nonliving seabed resource assessments for economic and management purposes. Out of 57 case studies, habitat mapping was intended to be part of an ongoing monitoring program in 24 cases, whereas the mapping was considered to be a one-off exercise in 33 cases. However, out of the 33 one-off cases, the authors considered that their habitat map would form the baseline for monitoring future changes in 24 cases. This suggests that governments and regulators generally view habitat mapping as a useful means of measuring and monitoring change. There appears to be a disconnect between the stated purpose of habitat mapping in most case studies (i.e., government management), and the end users and clients perceived by case study authors to be the main stakeholders. The data collected during a habitat mapping survey is composed of many types, including electronic data together with physical samples (sediments and biota) that must first be analyzed before numerical or classification type data are obtained.
The computation of slope and aspect angles for a cell is a common procedure in environmental studies and remote sensing applications in which topography is important. This article demonstrates that the slope/aspect angle derived from the neighboring elevation points best depicts the surface orientation for a larger cell. It is suggested that, rather than first resampling elevation datasets of a finer resolution to a larger cell size commensurate with other data in a study and then deriving slope/aspect angles, a mean slope/aspect angular measurement be derived directly from the higher resolution data for each larger cell size. -from Author
A software tool is described for the extraction of geomorphometric land surface variables and features from Digital Elevation Models (DEMs). The ArcGeomorphometry Toolbox consists of a series of Python/Numpy processing functions, presented through an easy-to-use graphical menu for the widely used ArcGIS package. Although many GIS provide some operations for analysing DEMs, the methods are often only partially implemented and can be difficult to find and used effectively. Since the results of automated characterisation of landscapes from DEMs are influenced by the extent being considered, the resolution of the source DEM and the size of the kernel (analysis window) used for processing, we have developed a tool to allow GIS users to flexibly apply several multi-scale analysis methods to parameterise and classify a DEM into discrete land surface units. Users can control the threshold values for land surface classifications. The size of the processing kernel can be used to identify land surface features across a range of landscape scales. The pattern of land surface units from each attempt at classification is displayed immediately and can then be processed in the GIS alongside additional data that can assist with a visual assessment and comparison of a series of results. The functionality of the ArcGeomorphometry toolbox is described using an example DEM.