NAM precipitation #3096
NAM precipitation #3096
Conversation
Calculate cumulative precip
Fix 00 and 12 hour accumulation interval
Calculate cumulative precip
Fix 00 and 12 hour accumulation interval
…nto task/nam-precip/3082
api/pyproject.toml
Outdated
| @@ -58,7 +58,7 @@ rope = "^1" | |||
| jsonpickle = "^3.0.0" | |||
| pytest-watch = "^4.2.0" | |||
| pytest-testmon = "^2.0.0" | |||
| ruff = "^0.0.286" | |||
| ruff = "^0.0.287" | |||
There was a problem hiding this comment.
This is weird, main is already using ruff @ 0.0.287
There was a problem hiding this comment.
Likely due to a merge from main instead of rebasing. Same for poetry.lock and package.json, yarn.lock and CriticalHoursCell.
| # Iterate through all the predictions. | ||
| prev_prediction = None | ||
|
|
||
| # 00 and 12 hour model runs accumulate precipitation in 12 hour intervals, 06 and 18 hour accumlate in |
There was a problem hiding this comment.
Makes sense, just wondering if we want to add some tests since due to the complexity.
There was a problem hiding this comment.
Hmmm...I'll see if I can factor all this code out into a utility function that we can test. Unless you have another thought for how we might test?
Codecov Report
@@ Coverage Diff @@
## main #3096 +/- ##
==========================================
+ Coverage 84.73% 84.74% +0.01%
==========================================
Files 300 300
Lines 10118 10133 +15
Branches 591 591
==========================================
+ Hits 8573 8587 +14
- Misses 1383 1384 +1
Partials 162 162
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more 📢 Have feedback on the report? Share it here. |
api/app/jobs/model_data.md
Outdated
|
|
||
| ### To 12:00 PST (13:00 PDT) | ||
|
|
||
| Fire weather forecasters in the BC Wildfire Service create daily "noon forecasts", where 'noon' refers to 12:00 PST, but during the fire weather season, most of BC is on Daylight Savings Time, so 12:00 PST = 13:00 PDT = 20:00 UTC. As explained above in the [NOAA section](#noaa), there is no model data specifically for 20:00 UTC, so instead we fetch the data for 18:00 and 21:00 UTC, and then perform additional linear interpolation (on top of the interpolation done for weather station locations) to predict weather behaviour for 20:00 UTC. This means that the modelled weather values for 21:00 UTC are weighted twice as heavily as those for 18:00 UTC, since the former is twice as close to our target time of 20:00. | ||
|
|
||
| This time-based linear interpolation is done as part of the data analysis process performed when the model data is retrieved from our third-party sources. See `process_grib.py` | ||
|
|
||
| ### Calculating predicted 24 hour precipitation | ||
|
|
||
| For any given date, Morecast 2 displays the precipitation predicted to fall in the 24 hours previous to that date. For example, for August 30, 2023, the predicted precip covers a time period from August 29, 2023 20:00 UTC to August 30, 2023 20:00 UTC. The first 24 hours of a numerical weather model run present a challenge because there is no way to calculate the predicted 24 hour precip purely from the model run data. Consider a HRDPS model run on August 30 at 12:00 UTC. Morecast 2 needs to display the predicted precipitation for August 30 at 20:00, but at that point we only have 8 hours of accumulated precipitation (from 12:00 to 20:00) from our model run, we are potentially missing precipitation that could have fallen from Aug 29 20:00 to August 30 12:00 which means we're missing 16 hours of data. To work around this, we pull actual rainfall values for that 16 hour period from our hourly_actuals table which scrapes observed values fromthe WF1 API. |
There was a problem hiding this comment.
Where does this part happen? Pulling WF1 data from hourly_actuals
There was a problem hiding this comment.
We have a cron job that runs the .../api/app/jobs/hourly_actuals.py script to scrape values reported by weather stations via the WF1 API.
|
Kudos, SonarCloud Quality Gate passed!
|
Test Links:
Landing Page
MoreCast 2.0
Percentile Calculator
MoreCast
C-Haines
FireBat
FireBat bookmark
Auto Spatial Advisory (ASA)
HFI Calculator