Tropical Cyclone Wind Exposure for the Eastern Pacific 1900-2016

MarineCadastre.gov Data Registry

URL where the source data were originally accessed.

Provided Extended Best Track (EBT) storm points data used for calculating climatology buffers (MEWs).

URL where the source data were originally accessed.

Provided tropical cyclone storm segments data.

URL where the source data were originally accessed.

Contributed lease block and protraction diagram data and information.

CREATE THE GRID 1. Within the Pacific Ocean basin, download the OCS lease blocks as the grid by which exposure will be modeled. 2. Remove all unnecessary fields and modify attributes as needed.

CLIMATOLOGY STEPS (MEW calculations)

3. Using the Extended Best Track (EBT) dataset (NOTE: this is an Atlantic Ocean basin data set), calculate the Maximum Extent of Winds (MEW) values following these steps. MEW are wind radii distances per standard wind threshold (34kt, 50kt, 64kt).

4. From http://rammb.cira.colostate.edu/research/tropical_cyclones/tc_extended_best_track_dataset/data/ebtrk_atlc_1988_2016.txt, right-click and save the EBT data (1988-2016) as a text file and open in Excel, setting custom fixed-width columns within the import wizard. Insert missing column breaks where necessary and set the width carefully, capturing the appropriate characters per column. Save file.

5. Assign each column a field name based on metadata here: (http://rammb.cira.colostate.edu/research/tropical_cyclones/tc_extended_best_track_dataset/docs/ebtrk_readme.txt).

6. Remove all EBT features in which storm type equals L, W, or D. (NOTE: No storm type values of 'D' were noted in the EBT data, even though the designation is stated within the metadata.)

7. Remove -99 values within fields.

8. Remove features if wind radii values in all four quadrants are zero. Per feature, even if only one of the four quadrants (per wind threshold) has a value above zero, retain the zero values in the other three cells for that wind threshold.

9. Remove all features with maximum wind values below 34 knots. (By definition, no wind radii values exist for EBT features with maximum winds below 34 knots, so these features are not relevant for MEW calculations.)

10. Modify the "longitudeWest" column by converting the values to negative values, so that features will plot appropriately (i.e. in the eastern hemisphere). Where necessary, convert longitude values within the eastern hemisphere by subtracting the value from 360.

11. Add the modified EBT data to ArcMap and export table as feature class.

12. Perform Select By Location, where all EBT points occurring over land (using a global countries layer) are selected. Include a 150nm buffer of the countries layer (with small islands removed) to select those EBT points occurring over land. (This is based on tropical cyclone literature, from Kruk 2010.) Export these selected features as a new feature class. These features will be used to calculate the average MEW values (buffer distances) for overland storm track segments. Switch the selection to select those features over water only. Export these selected features as a new feature class. These features will be used to calculate the average MEW values (buffer distances) for overwater storm track segments. If any overwater points exist over the Pacific Ocean basin, remove these.

13. For the overwater feature class, generate (storm) "category" values in a new field, based on wind speed values (see https://www.nhc.noaa.gov/aboutsshws.php). Due to the low sample of Cat 5 storms (1988-2016 = 28 points), combine these storms' (EBT points) wind radii values with the Cat 4 points when calculating MEW.

14. For the overland feature class, generate (storm) "category" values in a new field, based on wind speed values (see https://www.nhc.noaa.gov/aboutsshws.php). Due to the low sample of Cat 5 storms (1988-2016 = 24 points), combine these storms' (EBT points) wind radii values with the Cat 4 points when calculating MEWs.

15. Calculate MEW values per category for the overwater and overland features (separately) and record in a table. Compute averages for right (northeast, southeast) and left (northwest, southwest) quadrants.

16. Multiply all values by 0.67 to compute buffers for Eastern Pacific (EPAC) storms. This calculation is based on literature that has compared tropical cyclone sizes across global basins. This literature notes that wind swaths of EPAC tropical storms are approximately 33% smaller than North Atlantic tropical storms. (NOTE: size proportions between the North Atlantic and Eastern Pacific storms are variable within the literature.)

STORM TRACKS STEPS

17. Extract those features in the IBTrACS data that are: 1) attributed to the Eastern Pacific (EP) basin; 2) occur in or after 1900; and 3) have maximum wind values above 33 knots.

18. Extract those features that are attributed as extratropical (ET), subtropical (SS), or tropical (TS). The Eastern Pacific only had one feature denoted at subtropical, so it was removed and not included in the exposure modeling.

19. Feature (segments) from EPAC that crossed the International Dateline were trimmed, as these features cause issues when running the buffer tool in subsequent steps.

20. Create a (storm) "Category" field. For features that are tropical storms ('TS'), modify the field value to the appropriate category, based on wind speed (see https://www.nhc.noaa.gov/aboutsshws.php). Also designate features as 'ET', 'SS', or 'TS' accordingly.

21. Separate the overland and overwater storm segments into separate feature classes. Do this by extracting all segments "completely contained" within the "land + 150nm buffer" layer. (NOTE: This was only done for the North Atlantic segments. The vast majority of Eastern Pacific segments (about 99%) within BOEM's Pacific lease blocks do not occur over land. Additionally, the MEW values for segments that do intersect land (i.e. Hawaiian Islands) compared to MEW values over water would be very minimal.)

22. Separate EPAC storm segments by the seven categories (SS, ET, TS, 1, 2, 3, 4/5) into separate layers. (NOTE: For the ET, SS, and TS subsets, make additional subsets of these for the following segments: a) segments =34-49kt; b) segments =50-63kt; and c) segments >=64kt. This is required since not every ET, SS, and TS segment has winds greater than 50kt or 64kt, like Cat 1-Cat 5 segments do by definition.)

MODELING EXPOSURE

23. Buffer each (Eastern Pacific) subset (Cat 1, Cat 2, etc.) based on the respective MEW calculations (34 data sets total). Set Side Type field to right or left, depending on the buffer being run. Merge respective wind speed buffers (34kt, 50kt, 64kt) and then dissolve by storm ID (Serial_Num), storm name (Name), and season (Season).

24. Perform Spatial Join of the composite grid and buffers (34kt, 50kt, 64kt) to generate wind threshold counts per grid cell.

25. Perform Spatial Join of the composite grid and IBTrACS storm segments to generate storm segment counts per block.

26. Join count values of all outputs (34kt, 50kt, 64kt, and segments) together, based on the gridID field.

27. Calculate an average "return interval" value (117 years / count) per grid cell for each wind threshold (34kt, 50kt, 64kt).

28. Remove all unnecessary fields and populate others according to data dictionary.