## DSM Workflows ## This wiki contains high level descriptions of workflows typically used to build density surface models (DSMs), some detailed descriptions showing how members of our workgroup have done it, and general questions about workflow. It is not intended to be a tutorial on all of the steps needed to build DSMs. Typically, a DSM is a two-stage model built from a line-transect or point-transect survey, in which the first stage models the detectability of animals or objects of interest by observers using detection functions, and the second stage models the spatial or spatiotemporal distribution of density, typically by relating observed counts along survey segments to covariates with a generalized addtive model or another correlative statistical framework. [Miller et al. (2013)][2] give a comprehensive overview of this process. Below, we reproduce their Figure 6, which shows these two stages as the two top green blocks. After the spatial model has been fitted, gridded maps of abundance (number of individuals per grid cell) may be predicted from gridded maps of the covariates, and uncertainty may be estimated.  **Figure 6** from [Miller et al. (2013)][2]: Flow diagram showing the modeling process for creating a density surface model. How to cite this FAQ -------------------- To obtain a citation, go to the [main project page](https://osf.io/5eza8) and click the Citation dropdown. Contact us ---------- To ask a question or suggest material for this FAQ, please contact [Jason Roberts](mailto:jason.roberts@duke.edu). [1]: https://osf.io/xsfdy/wiki [2]: https://doi.org/10.1111/2041-210X.12105 [3]: https://files.osf.io/v1/resources/xsfdy/providers/osfstorage/5cf69ad1fe9cf50018bd20cf?mode=render