Bryan M. Costa

Using Lidar Bathymetry and Boosted Regression Trees to Predict the Diversity and Abundance of Fish and Corals

Abstract 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.

Mapping Reef Fish and the Seascape: Using Acoustics and Spatial Modeling to Guide Coastal Management

Reef fish distributions are patchy in time and space with some coral reef habitats supporting higher densities (i.e., aggregations) of fish than others. Identifying and quantifying fish aggregations (particularly during spawning events) are often top priorities for coastal managers. However, the rapid mapping of these aggregations using conventional survey methods (e.g., non-technical SCUBA diving and remotely operated cameras) are limited by depth, visibility and time. Acoustic sensors (i.e., splitbeam and multibeam echosounders) are not constrained by these same limitations, and were used to concurrently map and quantify the location, density and size of reef fish along with seafloor structure in two, separate locations in the U.S. Virgin Islands. Reef fish aggregations were documented along the shelf edge, an ecologically important ecotone in the region. Fish were grouped into three classes according to body size, and relationships with the benthic seascape were modeled in one area using Boosted Regression Trees. These models were validated in a second area to test their predictive performance in locations where fish have not been mapped. Models predicting the density of large fish (≥29 cm) performed well (i.e., AUC = 0.77). Water depth and standard deviation of depth were the most influential predictors at two spatial scales (100 and 300 m). Models of small (≤11 cm) and medium (12–28 cm) fish performed poorly (i.e., AUC = 0.49 to 0.68) due to the high prevalence (45–79%) of smaller fish in both locations, and the unequal prevalence of smaller fish in the training and validation areas. Integrating acoustic sensors with spatial modeling offers a new and reliable approach to rapidly identify fish aggregations and to predict the density large fish in un-surveyed locations. This integrative approach will help coastal managers to prioritize sites, and focus their limited resources on areas that may be of higher conservation value.

Identifying Suitable Locations for Mesophotic Hard Corals Offshore of Maui, Hawai‘i

Mesophotic hard corals (MHC) are increasingly threatened by a growing number of anthropogenic stressors, including impacts from fishing, land-based sources of pollution, and ocean acidification. However, little is known about their geographic distributions (particularly around the Pacific islands) because it is logistically challenging and expensive to gather data in the 30 to 150 meter depth range where these organisms typically live. The goal of this study was to begin to fill this knowledge gap by modelling and predicting the spatial distribution of three genera of mesophotic hard corals offshore of Maui in the Main Hawaiian Islands. Maximum Entropy modeling software was used to create separate maps of predicted probability of occurrence and uncertainty for: (1) Leptoseris, (2) Montipora, and (3) Porites. Genera prevalence was derived from the in situ presence/absence data, and used to convert relative habitat suitability to probability of occurrence values. Approximately 1,300 georeferenced records of the occurrence of MHC, and 34 environmental predictors were used to train the model ensembles. Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) values were between 0.89 and 0.97, indicating excellent overall model performance. Mean uncertainty and mean absolute error for the spatial predictions ranged from 0.006% to 0.05% and 3.73% to 17.6%, respectively. Depth, distance from shore, euphotic depth (mean and standard deviation) and sea surface temperature (mean and standard deviation) were identified as the six most influential predictor variables for partitioning habitats among the three genera. MHC were concentrated between Hanaka‘ō‘ō and Papawai Points offshore of western Maui most likely because this area hosts warmer, clearer and calmer water conditions almost year round. While these predictions helped to fill some knowledge gaps offshore of Maui, many information gaps remain in the Hawaiian Archipelago and Pacific Islands. This approach may be used to identify other potentially suitable areas for MHCs, helping scientists and resource managers prioritize sites, and focus their limited resources on areas that may be of higher scientific or conservation value.

Linking Cetaceans to their Environment: Spatial Data Acquisition, Digital Processing and Predictive Modeling for Marine Spatial Planning in the Northwest Atlantic

Cetaceans are large bodied, long-lived and highly mobile marine animals that exhibit extensive migrations, as well as, high site fidelity in areas where they aggregate for feeding, socializing, mating or calving. The marine environment in which they live is characterized by complex spatial and temporal heterogeneity. Cetaceans respond to this dynamic spatial structure at a range of scales, as denoted by their space-use patterns (Kenney et al. 2001; Baumgartner & Mate 2005). Space-use patterns provide important information about distributions of cetaceans and resource managers need these patterns to develop targeted conservation policies and resource management strategies. Despite this urgent need, adequate, spatially-explicit datasets do not exist for many regions of the world. Often resource managers that are charged with protecting endangered or threatened cetaceans have to rely on datasets that are sparse in both space and time. In order to address these knowledge gaps, resource managers urgently require quantitative, spatially explicit data on cetacean species distributions and species—environment relationships at ecologically and operationally relevant scales.

Managers, modelers, and measuring the impact of species distribution model uncertainty on marine zoning decisions

Marine managers routinely use spatial data to make decisions about their marine environment. Uncertainty associated with this spatial data can have profound impacts on these management decisions and their projected outcomes. Recent advances in modeling techniques, including species distribution models (SDMs), make it easier to generate continuous maps showing the uncertainty associated with spatial predictions and maps. However, SDM predictions and maps can be complex and nuanced. This complexity makes their use challenging for non-technical managers, preventing them from having the best available information to make decisions. To help bridge these communication and information gaps, we developed maps to illustrate how SDMs and associated uncertainty can be translated into readily usable products for managers. We also explicitly described the potential impacts of uncertainty on marine zoning decisions. This approach was applied to a case study in Saipan Lagoon, Commonwealth of the Northern Mariana Islands (CNMI). Managers in Saipan are interested in minimizing the potential impacts of personal watercraft (e.g., jet skis) on staghorn Acropora (i.e., Acropora aspera, A. formosa, and A. pulchra), which is an important coral assemblage in the lagoon. We used a recently completed SDM for staghorn Acropora to develop maps showing the sensitivity of zoning options to three different prediction and three different uncertainty thresholds (nine combinations total). Our analysis showed that the amount of area and geographic location of predicted staghorn Acropora presence changed based on these nine combinations. These dramatically different spatial patterns would have significant zoning implications when considering where to exclude and/or allow jet skis operations inside the lagoon. They also show that different uncertainty thresholds may lead managers to markedly different conclusions and courses of action. Defining acceptable levels of uncertainty upfront is critical for ensuring that managers can make more informed decisions, meet their marine resource goals and generate favorable outcomes for their stakeholders.

Marine biogeographic assessment of the main Hawaiian Islands : a collaborative investigation

An understanding of the distribution of marine benthic habitats and associated biota in the Main Hawaiian Islands (MHI) is necessary in order to assess potential direct and indirect effects of renewable energy development. Benthic habitats in the MHI can be divided into three broad categories based on their depth: shallow (<30 m), mesophotic (30-150 m) and deep (>150 m). Shallow-coral reefs provide numerous natural and economic benefits to the state’s economy and are much better studied than mesophotic and deep-water coral reefs. Approximately 75 percent of the shallow-water (<30 m) area around the MHI has been characterized using satellite imagery, although the percentage varies by island, with less area mapped around Hawaiʻi and the windward sides of Maui and Kahoʻolawe. Seventeen datasets from shallow reef monitoring programs were compiled into a standardized database of benthic cover. A qualitative assessment of the data indicates that percent cover of major benthic taxonomic groups (e.g., live coral, macroalgae) varies at both the island and local scales, with coral cover generally lower around the most northwestern islands. Recently published spatial predictive models of mesophotic hard coral distributions in the ʻAuʻau Channel provided maps of probability of occurrence for Leptoseris spp., Montipora spp. and Porites spp. These models were created using presence and absence records for these genera and a suite of environmental predictor variables. Probability of occurrence for mesophotic hard corals was highest in the warmer, clearer, and calmer waters off the western coast of Maui between Hanakaoʻo Point and Papawai Point. Although less data is available in deeper habitats, a variety of deep-sea corals (DSC) have been documented in the Hawaiian Archipelago. Using presence-only data and a suite of environmental predictor variables, spatial predictive models were created for eighteen DSC groups to identify areas most likely to contain deep-sea coral habitat around the MHI. The distributions of DSC presence records varied among groups; however, records were often concentrated in particular locations, such as Cross Seamount, Makapuʻu Point, Makalawena Bank, Lō‘ihi Seamount, and the southern edge of Penguin Bank. Areas predicted to contain highly suitable DSC habitat broadly aligned with the locations of DSC presence records. The environmental variables of depth, distance to shore and slope were consistently the most important predictors across all models. For both mesophotic corals and DSC, model results can be used to guide future exploration and research, particularly in areas where few records exist. 1 NOAA National Centers for Coastal Ocean Science, Biogeography Branch, Silver Spring, MD, U.S.A. 3 CSS-Dynamac, Fairfax, VA, U.S.A. 6 NOAA Office of National Marine Sanctuaries, Pahānaumokuākea Marine National Monument, Honolulu, HI, U.S.A. 7 NOAA Pacific Islands Fisheries Science Center, Protected Species Division, Honolulu, HI, U.S.A. 8 University of Hawaiʻi at Mānoa, Fisheries Ecology Research Lab, Hawaiʻi, U.S.A. Photo credit: Kurt Kawamoto (NOAA NMFS/PIFSC)