The accuracy of spatially continuous environmental data, usually generated from point samples using spatial prediction methods, is crucial for evidence-informed environmental management and conservation. Improving the accuracy by identifying the most accurate methods is essential, but also challenging since the accuracy is often data specific and affected by multiple factors. Recently developed hybrid methods of machine learning methods and geostatistics have shown their advantages in spatial predictive modelling in environmental sciences and significantly improved predictive accuracy. An R package, ‘spm: Spatial Predictive Modelling’, has been developed to introduce these methods and has been recently released for R users. It not only introduces the hybrid methods for improving predictive accuracy, but can also be used to improve modelling efficiency. This presentation will briefly introduce the developmental history of novel hybrid geostatistical and machine learning methods in spm. It will introduce spm, by covering: 1) spatial predictive methods, 2) new hybrid methods of geostatistical and machine learning methods, 3) assessment of predictive accuracy, 4) applications of spatial predictive models, and 5) relevant functions in spm. It will then demonstrate how to apply some functions in spm to relevant datasets and to show the resultant improvement in predictive accuracy and modelling efficiency. Although in this presentation, spm is applied to data in environmental sciences, it can be applied to data in other relevant disciplines. Presentation at the 2018 useR! conference

<div>This data product contains geospatial seabed morphology and geomorphology information for the Beagle Marine Park and is intended for use by marine park managers, regulators, the general public and other stakeholders. A nationally consistent two-part (two-step) seabed geomorphology classification system was used to map and classify the distribution of key seabed features. </div><div><br></div><div>In step 1, semi-automated GIS mapping tools (GA-SaMMT; Huang et al., 2022; eCat Record 146832) were applied to bathymetry digital elevation models (DEM) in a GIS environment (ESRI ArcGIS Pro) to map polygon extents (topographic high, low, and planar) and quantitatively characterise their geometries. The geometric attributes were then used to classify each shape into discrete Morphology Feature types (Part 1: Dove et al., 2020; eCat Record 144305). In step 2, the seabed geomorphology was interpreted by applying additional datasets and domain knowledge to inform their geomorphic characterisation (Part 2: Nanson et al., 2023; eCat Record 147818). Where available, backscatter intensity, seabed imagery, seabed sediment samples and sub-bottom profiles supplemented the bathymetry DEM and morphology classifications to inform the geomorphic interpretations.</div><div><br></div><div>The Beagle Marine Park seabed morphology and geomorphology features were informed by a post survey report (Barrett et al., 2021). Seabed units were classified at multiple resolutions that were informed by the underlying bathymetry: </div><div><br></div><div>· A broad scale layer represents features that were derived from a 30 m horizontal resolution compilation DEM (Beaman et al 2022; eCat Record 147043). </div><div>· A series of medium and fine scale feature layers were derived from individual 1 m horizontal resolution DEMs (Nichol et al., 2019; eCat Record 130301). </div><div><br></div><div>The data product and application schema are fully described in the accompanying Data Product Specification. </div><div><br></div><div><em>Barrett, N, Monk, J., Nichol, S., Falster, G., Carroll, A., Siwabessy, J., Deane, A., Nanson, R., Picard, K., Dando, N., Hulls, J., and Evans, H. (2021). Beagle Marine Park Post Survey Report: South-east Marine Parks Network. Report to the National Environmental Science Program, Marine Biodiversity Hub. University of Tasmania.</em></div><div><br></div><div><em>Beaman, R.J. (2022). High-resolution depth model for the Bass Strait -30 m. <a href=https://dx.doi.org/10.26186/147043>https://dx.doi.org/10.26186/147043</a>, GA eCat Record 147043.&nbsp;</em></div><div><br></div><div><em>Dove, D., Nanson, R., Bjarnadóttir, L. R., Guinan, J., Gafeira, J., Post, A., Dolan, Margaret F.J., Stewart, H., Arosio, R., Scott, G. (2020). A two-part seabed geomorphology classification scheme (v.2); Part 1: morphology features glossary. Zenodo. <a href=https://doi.org/10.5281/zenodo.40752483>https://doi.org/10.5281/zenodo.4075248</a>; GA eCat Record 144305&nbsp;</em></div><div><br></div><div><em>Huang, Z., Nanson, R. and Nichol, S. (2022). Geoscience Australia's Semi-automated Morphological Mapping Tools (GA-SaMMT) for Seabed Characterisation. Geoscience Australia, Canberra. <a href=https://dx.doi.org/10.26186/146832>https://dx.doi.org/10.26186/146832</a>; GA eCat Record 146832 </em></div><div><em>&nbsp;</em></div><div><em>Nanson, R., Arosio, R., Gafeira, J., McNeil, M., Dove, D., Bjarnadóttir, L., Dolan, M., Guinan, J., Post, A., Webb, J., Nichol, S. (2023). A two-part seabed geomorphology classification scheme; Part 2: Geomorphology classification framework and glossary (Version 1.0) (1.0). Zenodo.<a href=https://doi.org/10.5281/zenodo.7804019>https://doi.org/10.5281/zenodo.7804019</a>; GA eCat Record 147818&nbsp;</em></div>

<div>This data product contains geospatial seabed morphology and geomorphology information for Flinders Reefs and Cairns Seamount (Coral Sea Marine Park). These maps are intended for use by marine park managers, regulators, the general public and other stakeholders. A nationally consistent two-part (two-step) seabed geomorphology classification system was used to map and classify the distribution of key seabed features. </div><div><br></div><div>In step 1, semi-automated GIS mapping tools (GA-SaMMT; Huang et al., 2022; eCat Record 146832) were applied to a bathymetry digital elevation model (DEM) in a GIS environment (ESRI ArcGIS Pro) to map polygon extents (topographic high, low, and planar) and to quantitatively characterise their geometries. Their geometric attributes were then used to classify each shape into discrete Morphology Feature types (Part 1: Dove et al., 2020; eCat Record 144305). In step 2, the seabed geomorphology was interpreted by applying additional datasets and domain knowledge to inform their geomorphic characterisation (Part 2: Nanson et al., 2023; eCat Record 147818). Where available, backscatter intensity, seabed imagery, seabed sediment samples and sub-bottom profiles supplemented the bathymetry DEM and morphology classifications to inform the geomorphic interpretations.</div><div><br></div><div>The Flinders Reefs seabed morphology and geomorphology maps were derived from an 8 m horizontal resolution bathymetry DEM compiled from multibeam surveys (FK200429/GA4861: Beaman et al., 2020; FK200802/GA0365: Brooke et al, 2020), Laser Airborne Depth Sounder (LADS), Light Detection and Ranging (LiDAR) and bathymetry supplied by the Australian Hydrographic Office.</div><div><br></div><div>A subset of the FK200802/GA0365 multibeam survey was gridded at 1 m horizontal resolution to derive the key morphology and geomorphology features at the top of Cairns Seamount (-35 to -66 m; within the upper mesophotic zone).</div><div><br></div><div>The data product and application schema are fully described in the accompanying Data Product Specification. </div><div><br></div><div><em>Beaman, R., Duncan, P., Smith, D., Rais, K., Siwabessy, P.J.W., Spinoccia, M. 2020. Visioning the Coral Sea Marine Park bathymetry survey (FK200429/GA4861). Geoscience Australia, Canberra. <a href=https://dx.doi.org/10.26186/140048>https://dx.doi.org/10.26186/140048</a>; GA eCat record 140048</em></div><div><br></div><div><em>Brooke, B., Nichol, S., Beaman, R. 2020. Seamounts, Canyons and Reefs of the Coral Sea bathymetry survey (FK200802/GA0365). Geoscience Australia, Canberra. <a href=https://dx.doi.org/10.26186/144385>https://dx.doi.org/10.26186/144385</a>; GA eCat record 144385</em></div><div><br></div><div><em>Dove, D., Nanson, R., Bjarnadóttir, L. R., Guinan, J., Gafeira, J., Post, A., Dolan, Margaret F.J., Stewart, H., Arosio, R., Scott, G. (2020). A two-part seabed geomorphology classification scheme (v.2); Part 1: morphology features glossary. Zenodo. <a href=https://doi.org/10.5281/zenodo.4075248>https://doi.org/10.5281/zenodo.4075248</a>; GA eCat Record 144305 </em></div><div><br></div><div><em>Huang, Z., Nanson, R. and Nichol, S. (2022). Geoscience Australia's Semi-automated Morphological Mapping Tools (GA-SaMMT) for Seabed Characterisation. Geoscience Australia, Canberra. <a href=https://dx.doi.org/10.26186/146832>https://dx.doi.org/10.26186/146832</a>; GA eCat Record 146832</em></div><div><br></div><div><em>Nanson, R., Arosio, R., Gafeira, J., McNeil, M., Dove, D., Bjarnadóttir, L., Dolan, M., Guinan, J., Post, A., Webb, J., Nichol, S. (2023). A two-part seabed geomorphology classification scheme; Part 2: Geomorphology classification framework and glossary (Version 1.0) (1.0). Zenodo. <a href=https://doi.org/10.5281/zenodo.7804019>https://doi.org/10.5281/zenodo.7804019</a>; GA eCat Record 147818 </em></div>

Spatial predictive models have been increasingly employed to generate spatial predictions for environmental management and conservation in parallel to the advancement in data acquisition, data processing and computing capabilities. The accuracy of predictive models and their predictions is crucial to evidence-informed decision making and policy. However, the accuracy of predictive models in general is unknown and often accessed using error measures or even correlation measure. In this study, we clarified relevant issues about variance explained for predictive models (VEcv), established the relationships between commonly used predictive error measures like root mean square error (RMSE) and VEcv, unified these measures under VEcv, discovered that VEcv is independent of unit/scale and data variation, quantified the relationships between these error measures and data variation, and quantified the relationship between relative root mean square error (RRMSE) and relative mean absolute error (RMAE). We then assessed the performance of predictive models in the environmental sciences based on about 300 previously published applications and then classified the predictive models based on their performance. This study provided a tool to directly compare the accuracy of predictive models for data with different unit/scale and variation, and established a cross-disciplinary context and benchmark for assessing predictive models in environmental sciences and other disciplines. Recommendations for future studies were provided to objectively assess the performance of predictive models and make the accuracy of predictive models for different disciplines directly comparable. Abstract presented at the 23rd Australian Statistical Conference 2016

1. A robust scientific conclusion is the result of a rigorous scientific process. In observational ecology, this process involves making inferences about a population from 20 a sample. The sample is crucial, and is the result of implementing a survey design. A good survey design ensures that the data from the survey is capable of answering the research question. An even better design, such as spatially balanced designs, will also aim to reduce uncertainty as far as budgets will allow. 2. In many study areas, there are `legacy sites', that already have a time-series observed, and return visits to these sites are beneficial to enhance examination of temporal variability. We propose a method to incorporate these legacy sites into the survey effort whilst also maintaining spatial balance. This is the first formal method to perform this task. 3. Simulation experiments indicate that incorporating the spatial location of legacy sites increases spatial balance and decreases uncertainty in inferences (smaller standard errors in mean estimates). We illustrate the process using a proposed survey of a large marine reserve in South-Eastern Australia, where quantification of the reserve's biodiversity is required. 4. Our approach allows for integration of legacy sites into a new spatially-balanced 35 design, increasing efficiency. Scientists, managers and funders alike will benefit from this methodology { it provides a tool to provide efficient survey designs around established ones. In this way, it can aid integrated monitoring programs. An R-package that implements these methods, called MBHdesign, is available from CRAN. <b>Citation:</b> Foster, S.D., Hosack, G.R., Lawrence, E., Przeslawski, R., Hedge, P., Caley, M.J., Barrett, N.S., Williams, A., Li, J., Lynch, T., Dambacher, J.M., Sweatman, H.P.A. and Hayes, K.R. (2017), Spatially balanced designs that incorporate legacy sites. <i>Methods Ecol Evol</i>, 8: 1433-1442. https://doi.org/10.1111/2041-210X.12782

Spatial predictive methods are increasingly being used to generate predictions across various disciplines in environmental sciences. Accuracy of the predictions is critical as they form the basis for environmental management and conservation. Therefore, improving the accuracy by selecting an appropriate method and then developing the most accurate predictive model(s) is essential. However, it is challenging to select an appropriate method and find the most accurate predictive model for a given dataset due to many aspects and multiple factors involved in the modeling process. Many previous studies considered only a portion of these aspects and factors, often leading to sub-optimal or even misleading predictive models. This study evaluates a spatial predictive modeling process, and identifies nine major components for spatial predictive modeling. Each of these nine components is then reviewed, and guidelines for selecting and applying relevant components and developing accurate predictive models are provided. Finally, reproducible examples using spm, an R package, are provided to demonstrate how to select and develop predictive models using machine learning, geostatistics, and their hybrid methods according to predictive accuracy for spatial predictive modeling; reproducible examples are also provided to generate and visualize spatial predictions in environmental sciences. <b>Citation:</b> Li, J. A Critical Review of Spatial Predictive Modeling Process in Environmental Sciences with Reproducible Examples in R. Appl. Sci. 2019, 9, 2048. https://doi.org/10.3390/app9102048

Seabed sediment predictions at regional and national scales in Australia are mainly based on bathymetry-related variables due to the lack of backscatter-derived data. In this study, we applied random forests (RFs), hybrid methods of RF and geostatistics, and generalized boosted regression modelling (GBM), to seabed sand content point data and acoustic multibeam data and their derived variables, to develop an accurate model to predict seabed sand content at a local scale. We also addressed relevant issues with variable selection. It was found that: (1) backscatter-related variables are more important than bathymetry-related variables for sand predictive modelling; (2) the inclusion of highly correlated predictors can improve predictive accuracy; (3) the rank orders of averaged variable importance (AVI) and accuracy contribution change with input predictors for RF and are not necessarily matched; (4) a knowledge-informed AVI method (KIAVI2) is recommended for RF; (5) the hybrid methods and their averaging can significantly improve predictive accuracy and are recommended; (6) relationships between sand and predictors are non-linear; and (7) variable selection methods for GBM need further study. Accuracy-improved predictions of sand content are generated at high resolution, which provide important baseline information for environmental management and conservation. <b>Citation:</b> Li, J.; Siwabessy, J.; Huang, Z.; Nichol, S. Developing an Optimal Spatial Predictive Model for Seabed Sand Content Using Machine Learning, Geostatistics, and Their Hybrid Methods. <i>Geosciences</i> 2019, 9, 180. https://doi.org/10.3390/geosciences9040180

The seascape of the vast Australian continental margin is characterised by numerous submarine canyons that represent an equally vast array of geomorphic and oceanographic heterogeneity. Theoretically, this heterogeneity translates into habitats that may vary equally widely in their ecological characteristics. Here we describe the methodology to develop a framework to broadly derive estimates of potential habitat condition (¿suitability¿ sensu lato) for pelagic and epibenthic megafauna (including demersal fishes), and benthic infauna in all of Australia¿s known submarine canyons. We find that the high geomorphic and oceanographic diversity of submarine canyons creates a multitude of potential habitat types. In general, it appears that canyons may be particularly high-quality for benthic species. Canyons that incise the shelf tend to score higher in habitat potential than those confined to the slope. Canyons with particularly high habitat potential are located mainly off the Great Barrier Reef, the NSW coast, the eastern margin of Tasmania and Bass Strait, and on the southern margin. Many of these canyons have complex bottom topography, are likely to be productive, and have less intense sediment disturbance regimes. The framework presented here can be relevant ¿ once refined and comprehensively validated with ecological data - in a management and conservation context to identify canyons (or groups of canyons) that are likely to represent high-value habitat along a vast continental margin where marine planning decisions may require spatial prioritization decisions. <b>Citation:</b> Zhi Huang, Thomas A. Schlacher, Scott Nichol, Alan Williams, Franziska Althaus, Rudy Kloser, A conceptual surrogacy framework to evaluate the habitat potential of submarine canyons, <i>Progress in Oceanography</i>, Volume 169, 2018, Pages 199-213, ISSN 0079-6611, https://doi.org/10.1016/j.pocean.2017.11.007

Rapid, efficient, and accurate prediction of mineral occurrence that takes uncertainty into 20 account is essential to optimise defining exploration targets. Traditional approaches to mineral 21 potential mapping often fail to fully appreciate spatial uncertainties of input predictors and their 22 spatial cross-correlation. In this study a stochastic technique based on multivariate 23 geostatistical simulations and ensemble tree-based learners is introduced for predicting and 24 uncertainty quantification of mineral exploration targets. The technique is tested on a synthetic 25 case inspired by the characteristics of a hydrothermal mineral system model and a real-world 26 dataset from the Yilgarn Craton in Western Australia. Results from the two cases proved the 27 superior performance and robustness of the proposed stochastic technique, especially when 28 dealing with high dimensional and large data sets. <b>Citation:</b> Talebi, H., Mueller, U., Peeters, L.J.M. et al. Stochastic Modelling of Mineral Exploration Targets. <i>Math Geosci </i>54, 593–621 (2022). https://doi.org/10.1007/s11004-021-09989-z

Spatial distribution of sponge species richness (SSR) and its relationship with environment are important for marine ecosystem management, but they are either unavailable or unknown. Hence we applied random forest (RF), generalised linear model (GLM) and their hybrid methods with geostatistical techniques to SSR data by addressing relevant issues with variable selection and model selection. It was found that: 1) of five variable selection methods, one is suitable for selecting optimal RF predictive models; 2) traditional model selection methods are unsuitable for identifying GLM predictive models and joint application of RF and AIC can select accuracy-improved models; 3) highly correlated predictors may improve RF predictive accuracy; 4) hybrid methods for RF can accurately predict count data; and 5) effects of model averaging are method-dependent. This study depicted the non-linear relationships of SSR and predictors, generated spatial distribution of SSR with high accuracy and revealed the association of high SSR with hard seabed features. <b>Citation:</b> Jin Li, Belinda Alvarez, Justy Siwabessy, Maggie Tran, Zhi Huang, Rachel Przeslawski, Lynda Radke, Floyd Howard, Scott Nichol, Application of random forest, generalised linear model and their hybrid methods with geostatistical techniques to count data: Predicting sponge species richness, <i>Environmental Modelling & Software</i>, Volume 97, 2017, Pages 112-129, https://doi.org/10.1016/j.envsoft.2017.07.016