Olympic Peninsula Regional Anadromous Fish Habitat Intrinsic Potential Modeling

2015 Region Wide IP Modeling

Outer Coast IP Map Portal

Previous IP Modeling Work

Phase II Models

IP Review Results / Documents

Phase I Models


Wild Salmon Center




PDF Document Phase II Report February 2013

PDF Document Phase I Report June 2012


Marine staff
Field Operations

A discussion on Binning / Symbolization of IP Scores

Currently, this work is now in its second phase, with region wide IP models being applied, mapped, and provided here for fish biologists and local experts to review. An important part of this work also includes obtaining a consensus concerning where to set biologically significant breakpoints in order to produce a consistent symbolization scheme for each species.

Toward that end, a discussion of how to represent IP scores on maps using binning schemes as a symolization method is presented here.

Problems with correlation on the ground:

The aim of any representation on a map is to provide a sense of what is actually on the ground in the area depicted. More specifically, how does one represent IP scores for each species in a way that reflects where the fish are most likely to be found, particularly in areas that are unmodified by anthropogenic influences. As a concept, this is easy to understand, but in reality, it is very difficult to do:

Regionality is a strong influence; In the Puget Sound region, the presence of five fish might be considered an indicator of highly suitable habitat, and be depicted by a high IP score. In areas such as the Outer Washington Coast, less affected by anthropogenic influence, the presence of five fish might be considered an indicator of more marginal habitat.

Establishing a correlation between what the fish are doing on the ground and what the IP scores depict is much harder for the OWC region. Good field data for presence/absence and spawning locations does not exist in an easily useable GIS form, rendering any effort to get good correlation between IP scores on the NetMap reaches and these data a difficult, time consuming task at best.[1]

Some of the issues to consider when discussing data available for correlative purposes[2]:

Binning in published literature

A wide variety of binning schemes have been used by other authors, notably a 9-bin equal interval [Agrawal et al, 2005], a five bin equal interval [Busch et al, 2011] and a variety of other symbolization schemes.

Of significance is the categorization of "High" IP scores:

Binning / Symbolization for Phase II IP Maps

Note that the above authors undertook studies in areas other than the OWC, so the following adaptations for the purpose of symbolizing OWC should be taken with the caveat that they may not be appropriate for salmonid species here:

Having set the bin for "High" IP scores, the next step was to divide the remainder of the scores above zero into two equal sized bins appropriate for each species. This required taking the mean value for each of the histogram distributions relevant to each species and placing the break point between "Medium" and "Low" there. Zero values are represented separately. For Coho, winter and summer rearing scores were treated separately, with 0.75 assigned to the "High" bin for both [see above]. Each histogram is provided here:

Zero IP values are rendered separately in all symbolizations. It should be noted that there is considerable interaction with the NetMap model in regards to IP values. This is because for some purposes, a parameter value relevant to the life cycle being modeled (Chinook spawning only model) is considered a "drop dead" limit even though the fish bearing network may be more extensive.

[1] Mindi Sheer has commented that she had put in a lot of time preparing existing data but it is a very large job and has not been completed. [Busch 2011, Mindi Sheer pers comment, Nov 2012]

[2] [Busch 2011, Mindi Sheer pers comment, Nov 2012]