Predictive mapping of the biotic condition of conterminous U.S. rivers and streams.

Understanding and mapping the spatial variation in stream biological condition could provide an important tool for conservation, assessment, and restoration of stream ecosystems. The USEPA's 2008-2009 National Rivers and Streams Assessment (NRSA) summarizes the percentage of stream lengths within the conterminous United States that are in good, fair, or poor biological condition based on a multimetric index of benthic invertebrate assemblages. However, condition is usually summarized at regional or national scales, and these assessments do not provide substantial insight into the spatial distribution of conditions at unsampled locations. We used random forests to model and predict the probable condition of several million kilometers of streams across the conterminous United States based on nearby and upstream landscape features, including human-related alterations to watersheds. To do so, we linked NRSA sample sites to the USEPA's StreamCat Dataset; a database of several hundred landscape metrics for all 1:100,000-scale streams and their associated watersheds within the conterminous United States. The StreamCat data provided geospatial indicators of nearby and upstream land use, land cover, climate, and other landscape features for modeling. Nationally, the model correctly predicted the biological condition class of 75% of NRSA sites. Although model evaluations suggested good discrimination among condition classes, we present maps as predicted probabilities of good condition, given upstream and nearby landscape settings. Inversely, the maps can be interpreted as the probability of a stream being in poor condition, given human-related watershed alterations. These predictions are available for download from the USEPA's StreamCat website. Finally, we illustrate how these predictions could be used to prioritize streams for conservation or restoration.

Assessing the accuracy and stability of variable selection methods for random forest modeling in ecology.

Random forest (RF) modeling has emerged as an important statistical learning method in ecology due to its exceptional predictive performance. However, for large and complex ecological data sets, there is limited guidance on variable selection methods for RF modeling. Typically, either a preselected set of predictor variables are used or stepwise procedures are employed which iteratively remove variables according to their importance measures. This paper investigates the application of variable selection methods to RF models for predicting probable biological stream condition. Our motivating data set consists of the good/poor condition of n = 1365 stream survey sites from the 2008/2009 National Rivers and Stream Assessment, and a large set (p = 212) of landscape features from the StreamCat data set as potential predictors. We compare two types of RF models: a full variable set model with all 212 predictors and a reduced variable set model selected using a backward elimination approach. We assess model accuracy using RF's internal out-of-bag estimate, and a cross-validation procedure with validation folds external to the variable selection process. We also assess the stability of the spatial predictions generated by the RF models to changes in the number of predictors and argue that model selection needs to consider both accuracy and stability. The results suggest that RF modeling is robust to the inclusion of many variables of moderate to low importance. We found no substantial improvement in cross-validated accuracy as a result of variable reduction. Moreover, the backward elimination procedure tended to select too few variables and exhibited numerous issues such as upwardly biased out-of-bag accuracy estimates and instabilities in the spatial predictions. We use simulations to further support and generalize results from the analysis of real data. A main purpose of this work is to elucidate issues of model selection bias and instability to ecologists interested in using RF to develop predictive models with large environmental data sets.

Comparing spatial regression to random forests for large environmental data sets.

Environmental data may be "large" due to number of records, number of covariates, or both. Random forests has a reputation for good predictive performance when using many covariates with nonlinear relationships, whereas spatial regression, when using reduced rank methods, has a reputation for good predictive performance when using many records that are spatially autocorrelated. In this study, we compare these two techniques using a data set containing the macroinvertebrate multimetric index (MMI) at 1859 stream sites with over 200 landscape covariates. A primary application is mapping MMI predictions and prediction errors at 1.1 million perennial stream reaches across the conterminous United States. For the spatial regression model, we develop a novel transformation procedure that estimates Box-Cox transformations to linearize covariate relationships and handles possibly zero-inflated covariates. We find that the spatial regression model with transformations, and a subsequent selection of significant covariates, has cross-validation performance comparable to random forests. We also find that prediction interval coverage is close to nominal for each method, but that spatial regression prediction intervals tend to be narrower and have less variability than quantile regression forest prediction intervals. A simulation study is used to generalize results and clarify advantages of each modeling approach.

Estimating the Distribution of Harvested Estuarine Bivalves with Natural-History-Based Habitat Suitability Models.

Habitat suitability models are useful to estimate the potential distribution of a species of interest, particularly in the case of infaunal bivalves. Sampling for these bivalves is time- and cost-intensive, which is increasingly difficult for organizations or agencies that are limited by personnel and funds. Consequently, we developed a framework to identify suitable bivalve habitat in estuaries (FISBHE) - a habitat suitability index (HSI) modeling framework for NE Pacific estuaries that was parameterized with published natural-hi story information and existing habitat datasets, without requiring extensive field sampling of bivalves. Spatially explicit, rule-based habitat suitability models were constructed in a GIS for five species of bay-clams (Clinocardium nuttallii, My a arenaria, Tresus capax, Saxidomus gigantea, and Leukoma staminea) that are popular targets for recreational and commercial harvest in estuaries of the U.S. Pacific Northwest. Habitat rasters were produced for Yaquina and Tillamook estuaries (Oregon, USA) using environmental data (bathymetric depth, sediment % silt-clay, wet-season salinity, and burrowing shrimp presence/absence) from multiple studies (1953-2015). These habitat rasters then served as inputs in the final model which produced HSI classes ranging from 0-4 (lowest to highest suitability), dependent upon the number of habitat variables that fell within the sensitivity limits for each species of bay-clam. The models were tested with validation analyses and bay-clam occurrence data (reported in benthic community studies, 1996-2012) within each HSI class; logistic regression and Kendall's correlation coefficient both showed correspondence between predicted HSI classes and bay-clam presence/absence. Results also showed that the greatest presence probabilities occurred within habitats of highest predicted suitability, with the exception of M. arenaria in Tillamook Bay. The advantage of FISBHE is that disparate, independent sets of existing data are sufficient to parameterize the models, as well as produce and validate maps of habitat suitability. This approach can be transferred to data-poor systems with modest investment, which can be useful for prioritizing estuarine land-use decisions and could be used to estimate the vulnerability of this valued ecosystem good to changes in habitat quality and distribution.