Markov Chain Update. GIS Coastline Rasterize Coastline in Matlab.

Markov Chain Update

GIS Coastline

Rasterize Coastline in Matlab

Habitat = Spiny Lobster

Split up Patches

What we have:

A rasterized coastline based on a real GIS map

Patches that are based on the suitable habitat for Spiny Lobster geometric/flow structures (e.g. the north side of an island will be a different patch to the south side)

• Now – setup a Markov particle tracking simulation with a spatially explicit flow field (provided by Michael, from Sea Surface Height measurements)…

Flow Field from SSH

Velocitym/s

… and v component

Velocitym/s

Lots of fiddling to get Markov script to ‘sample’ u and v

velocities from the flow field

New method for starting particles … Randomly seeded within each

patch

+ when they hit land they bounce better… to their last “ocean” positon

For example

• Michael kindly got two flow fields – Jan 99 and June 99.

• I used our new Markov technique to generate connectivity matrixes for these two ocean states.

• PLD = 20days, competency after 2 days.• Spiny Lobster habitat

…. Unreal, but worth a look.

The i and j patches

Total Count Connectivity Matrixes

JANUARY 1999

JUNE 1999

j

j

i

i

Area Normalized Connectivity Matrixes

JANUARY 1999

JUNE 1999

• So getting there, we now need better flow fields and biological parameters to proceed.

• Game over