In USEPA/LakeTrophicModelling: Package to reproduce Hollister et al. (2015) Modeling Lake Trophic State: A Randrom Forest Approach.

September 03, 2015 \newline

Debra Peters\newline Editor, Ecosphere\newline USDA ARS\newline Las Cruces, NM 88003\newline

Dear Dr. Peters,

Please accept for your review the submission, "Modelling Lake Trophic State: A Random Forest Approach" to be published in Ecosphere. In this article we detail our use of Random Forests to model chlorophyll a and trophic state in lakes across the United States. This research makes three important contributions. First, most prior studies of lake trophic state focus on a limited number of lakes that span local or regional extents. Our models are built for lakes that span the conterminous United States. Second, we compare models with and without in situ water quality data and find that models with only broadly available data still provide reliable predictions. This allows for predictions to be made in nearly every lakes in the U.S. Third, we utilize a data mining algorithm, Random Forests, to build our model. This approach provides accurate and unbiased predictions. Lastly, we are proponents of practicing open science and as such, use open source tools, open data, and open access publishing whenever possible. Towards that end all of the data and source code for this manuscript are availble as an R package via GitHub (https://github.com/USEPA/LakeTrophicModelling) and a pre-print of this manuscript has been submitted to the PeerJ Preprint server (https://peerj.com/preprints/1319/). We understand that according to ESA policy this does not count as prior publication.

Sincerely,

Jeffrey W. Hollister 401 782 9655 [e-mail]

W. Bryan Milstead [e-mail]

Betty J. Kreakie [e-mail]

USEPA/LakeTrophicModelling documentation built on Oct. 15, 2020, 4:13 p.m.