Innovation Sprint Ideas

Research Paper Idea:

MoreCast 2.0 Raw Model Data vs Bias Adjustments vs Fire Behaviour vs Weather Data as Observed from Stations
- Publish Raw Grib Data on AWS S3
- API work
- Innovation Sprint to come up w/ coding requirements and then start paper once 2024 season kicks in
- https://github.com/google-deepmind/graphcast

++++++++++++

PO proposals for an Innovation Sprint:

What we have access to:
SFMS Layers
Weather Data
Fire Data: a) Geo BC, WFS and b) WF1 API
Lightning: Scrum Master requested access from DH July 19

See what’s possible with: **Business Intelligence Dashboards**

**PO’s EMCR briefings:**

A) Communication + visualisation needs
a. Needs automation: Fires per week vs historical percentile from last 20ys + Patterns
b. Analogue years (similar fire behaviour in the past)
c. Size of Fires vs Lighting
d. Currently using Jaspersoft (see recording 1)
e. Spatial + Fire Data combined + real time info + historical data + warnings
f. Control chart analogy: Rate of Change + Disparate Sources: Cumulative Daily Severity Rating vs Climate Normals
g. (nurses taking the pulse + ability to read past charts of pulse)
i. Initial Attack Failure
j. Area Burned
k. Early Warning system (going over 6-10 hectares - when? Area burned per day?)
l. **** Need PO's code + static data dump: Long Term Weather Forecast Predictions (Climate Analogues + Fire History Data) *****

B) Stats for fires (Wildfire Activity Slide) + Comparisons

C) Lighting Strikes vs Holdover Fires vs Strikes in a specific area vs # of strikes in a high HFI area of concern
Currently using WF1 Sparky App FOP

D) PSU Comms: WF updates: anomaly of Fire Indices
Build up index + Drought codes + Winds + any other parameters: Display on UI if these exceed historical normal (Thresholds)
Ex: HEADS UP notification: This index is on upward trend OR outside normals

E) CHAINES + Advisories
a. Combining multiple sources of info + WHY this is happening
i. Build up index
ii. High HFI
iii. Hot dry wind
iv. Haines index high …. So: this result is observed
b. Wind + Fires out of control + High Haines index

F) Need surface wind speed x min RH (vapour pressure deficit)
a. Lower atmosphere vs smoke columns
i. Wind = 4 klm @ 10 m BUT
ii. Wind = 70 klm @ 150 m

G) Rewrite SFMS (in C + open sourced from Canadian Fire Service)

H) Red Hat OpenShift Data Science:
https://www.redhat.com/en/technologies/cloud-computing/openshift/openshift-data-science

I) Hackathon: ASA + Mapping PMTiles Solutions:
PMT Tiles Visualizations + analysis + locations + attributes
Fire points + underlying warning levels (Control chart idea): Advisory Level impacts

J) Exposing maps on Mobile/Cell phones?

Let’s step in TECH + see what’s possible

++++++++++++

Recording 1 - 18 minutes
https://bcgov.sharepoint.com/:v:/r/teams/01324/Shared%20Documents/General/Recordings/BCWS%20-%20Agile%20PSU%20%E2%80%93%20Innovation-Meeting-Recording-2.mp4?csf=1&web=1&e=Mr4l3g

Recording 2 - 5 minutes
https://bcgov.sharepoint.com/:v:/r/teams/01324/Shared%20Documents/General/Recordings/BCWS%20-%20Agile%20PSU%20%E2%80%93%20Innovation-Meeting-Recording-1.mp4?csf=1&web=1&e=k6PPLf

The issue names no repository files, tests, or entry points; start by choosing one proposal and documenting its coding requirements and data sources. Review the linked GraphCast project and the available SFMS, weather, fire, lightning, and WF1 data before scoping an API, dashboard, or research deliverable. Done is a single agreed, actionable scope with acceptance criteria.

Written by the indexing model from the issue text.