ohi.model.pressures.matrix: Ocean Health Index: Pressures Matrix Model
In bbest/ohigui: Ocean Health Index - Graphical User Interface

Description Usage Arguments Details Value See Also

The pressures matrix model function computes a pressures weighting matrix based on regional attributes per category.

1	ohi.model.pressures.matrix(alpha, beta, calc = "avg")

`alpha`	the weighting matrix of the form [category x pressure]. Each rank weight must be an integer between 0 and 3 inclusive, or NA.
`beta`	the aggregation matrix of the form [region_id x category] to collapse across each category.
`calc`	type of calculation, whether avg (default), mean (diff't from avg?) or presence (results in 1 or 0).

Given:

g is the goal or subgoal (e.g., AO, CW, LIV, ECO, ...),
i is the region (e.g., 1, 2, 3, ...),
j is the pressures layer or stressor (e.g., cc_acid, fp_art_lb, etc.).
k is the category (e.g., habitat, sector, product, etc.)

There may be a component k for a given goal g such that p_w(g,i,j,k) and w(g,i,j,k).

p_{w}(g,i,j,k) = w(g,i,j,k) * p(i,j)

In these cases where there is a component k for goal g, there's an additional aggregation or formula to calculate w(g,i,j) based on the core rank weight α(g,j,k) from the original pressures matrix (as written in Halpern et al. (2012)) and some region-specific data for each category k β(i,k).

This function ohi.model.pressures.matrix will aggregate a category-specific weighting matrix α(g,j,k) [category x pressure] using region-specific data β(g,i,k) into a [region_id x pressure] matrix w(g,i,j) used in ohi.model.pressures, such that:

w(g,i,j) = \frac{∑_k α({g,j,k}) * β({g,i,k})}{∑_k β({g,i,k})}

For the CP, CS goals, the weight depends on the extent A of habitat k in region i:

β(i,k) = A(i,k)
For the HAB goal, the weight depends on the presence of habitat k (i.e., if A(i,k)>0) in region i:

β(i,k) = hasHabitat(i,k)
For the LIV and ECO goals, the weight depends on the presence of sector k if data available for region i and sector k:

β(i,k) = hasSector(i,k)
For the NP goal, the weight depends on the peak dollar value of each product k across all years (see w_p from SI Equation S27) if data available for region i and product k:

β(i,k) = w_p(i,k)