Ten Years of Science Based on the Canadian Precipitation Analysis: A CaPA System Overview and Literature Review^†

† This paper is dedicated to the memory of Professor Peter Funder Rasmussen, an exceptional scientist, mentor, and friend who contributed significantly to the Canadian Precipitation Analysis project.

Figures & data

Table 1. Chronology of changes to the configuration of CaPA.

Version	Release Date	Configuration, Publications, and Changes from Previous Version
1.0	2004	Version used by Mahfouf et al. (2007) for the proof of concept
2.0	21 Sep. 2010	First candidate version for operational implementation Similar to version 2.2, but without the bias correction method to deal with the bias induced by the cubic root transform (Evans, 2013) Version used for the study by Carrera et al. (2010)
2.2	6 Apr. 2011	First operational implementation of CaPA–RDPA, providing both 6 h and 24 h precipitation estimates on a 15 km grid Early (available at T + 1 h) and late (available at T + 7 h) 6 h analyses provided A 24 h analysis valid at 1200 utc, available at T + 7 h, is also provided Version used for the performance evaluation presented in Lespinas et al. (2015) In Canada, CaPA assimilated data from ECCC’s weather stations as well as stations belonging to the Réseau Météorologique Coopératif du Québec (RMCQ) In the United States, CaPA assimilated data from the synoptic network, from American Meteorological Terminal Aviation Routine (METAR) and from the US Cooperative Observer Network (COOP) network
2.3	3 Oct. 2012	CaPA–RDPA relying on version 3.0.0 of RDPS to obtain its background field Increase in horizontal resolution from 15 km to 10 km.
2.4β8	6 May 2013	Minor improvements and bug fixes Beta 8 release of version 2.4 used to create a 2002–2012 reanalysis Although the reanalysis has a resolution of 10 km, the GEM model background field has a resolution of 24 km from Jan. 2002 until May 2004 and 15 km thereafter. It was interpolated onto a 10 km grid using a nearest neighbour algorithm
2.4	20 Aug. 2013	Minor improvements and bug fixes
3.0	18 Nov. 2014	CaPA relying on version 4.0.0 of the RDPS to obtain its background field Assimilation of QPEs from the Canadian weather radar network (31 radars) A preliminary 24 h analysis (available at T + 1 h) now included Version used by Fortin, Roy, et al. (2015) to assess the impact of radar data on the quality of the analysis
3.1	3 Jun. 2015	Quality-control algorithm improved in order to assimilate a higher number of solid precipitation observations
3.2	15 Dec. 2016	Minor improvements and bug fixes
4.0	9 Nov. 2016	QPEs from 33 American weather radars assimilated Observations of 0 mm inferred from METAR reports Observations from the British Columbia Wildfire Service and the Ontario Ministry of Natural Resources and Forest assimilated
4.1	2 Mar. 2018	Operational implementation of CaPA–HRDPA relying on version 4.4.0 of the HRDPS to obtain its background field Watershed averages of precipitation provided in shapefile format

Fig. 1 24 h precipitation total valid at 18 August 2017 1200 utc for (a) CaPA–RDPA and (b) CaPA–HRDPA.

Fig. 2 Enlargement of the Great Lakes region for 24 h precipitation total valid at 18 August 2017 1200 utc for (a) CaPA–RDPA, (b) CaPA–HRDPA, and (c) in situ and radar observations.

Fig. 3 Confidence index (CFIA) for the 24 h precipitation total valid at 18 August 2017 1200 utc generated by CaPA–HRDPA.

Fig. 4 Datasets used for evaluating the added value of the BC Forest network (purple dots, stations withheld for verification; black dots, stations belonging to the BC Forest network), evaluating the added value of radar data (white circles for the stations, blue circles for the radar footprint), and evaluating the added value of increasing the horizontal resolution of the analysis from 10 km to 2.5 km (red dots). Green dots correspond to the location of the remaining stations assimilated by CaPA.

Fig. 5 Impact of three changes to CaPA on the Frequency Bias Index (FBI_m, left column) and Equitable Threat Score (ETS, right column) of 24 h analyses for precipitation thresholds between 0.2 and 50 mm per 24 h (x-axis). First row: impact of assimilating BC Forest data, second row: impact of assimilating radar data, third row: impact of increasing horizontal resolution from 10 to 2.5 km. The blue line (with black squares) corresponds to the initial bias and skill before the change, and the red line (with round dots) corresponds to the bias and skill after the change. Differences are not statistically significant where the background of the graph is shaded (90% confidence level). Evaluation performed over the summer of 2017 (June, July, and August). Results are not shown when the number of events for a given threshold is less than 30.

Table 2. Studies dedicated to the evaluation of CaPA or multiple gridded precipitation datasets including CaPA.

Contribution	Area of Interest	Period of Study	Version of CaPA	Comparison to Precip. Obs.	Proxies Used for Indirect Comparisons	Other Datasets Evaluated	Preferred Product (where applicable)
Boluwade et al. (2018)	Lake Winnipeg basin (LWB)	2002–2005	2.4β8 reanalysis	Yes, dependent observations	Streamflow and WATFLOOD hydrological model	—	—
Bajamgnigni Gbambie et al. (2017)	province of Quebec	2002–2010	2.4β8 reanalysis	No	Streamflow and HSAMI^a hydrological model	ANUSPLIN^b, MDDELCC^c	CaPA, particularly in the central and northern parts of the province
Hanes et al. (2017)	northern Ontario	summers of 2014 and 2015	3.0	Yes, independent observations	Fire Weather Index	Thin-plate smoothing interpolation method	CaPA, particularly in regions with radar coverage
Wong et al. (2017)	Canada (results provided for each ecozone)	2002–2012	2.4β8 reanalysis	Yes, dependent observations	—	ANUSPLIN^b, Princeton, WFDEI^d [CRU^e], WFDEI^d [GPCC^f], NARR^g, PCIC^h, NA-CORDEX^i	WFDEI^d [GPCC^f] and CaPA
Lespinas et al. (2015)	Canada (results provided for each ecozone)	2009–2012	2.2	Yes, independent observations	—	—	—
Fortin, Roy, et al. (2015)	southern Canada (region with radar coverage)	summer of 2014	3.0	Yes, independent observations	—	—	—
Eum, Dibike, et al. (2014)	Athabaska river watershed	2002–2010	2.4β8 reanalysis	Yes, dependent observations	Streamflow and VIC^j hydrological model	ANUSPLIN^b, NARR^g	CaPA, post 2003 and for regions with sparse observational networks

^aService Hydrométéorologique Apports Modules Intermédiaires.

^bAustralian National University spline interpolation method.

^cMinistère du Développement Durable, de l’Environnement et de la Lutte au Changement Climatique.

^dWATCH-Forcing-Data-ERA-Interim.

^eClimate Research Unit.

^fGlobal Precipitation Climatology Centre.

^gNorth American Regional Reanalysis.

^hPacific Climate Information Consortium.

ⁱNorth American Coordinated Regional Climate Downscaling Experiment.

^jVariable Infiltration capacity.

Table 3. Methods for accessing CaPA products over the Internet.

Website	Products Available	Format	Time Period	Version of CaPA	Support
Government of Canada weather portal https://weather.gc.ca/analysis	CaPA–RDPA, 6 h and 24 h analyses	Static images in PNG format	Last 24 hours	Latest operational version	Operational support by MSC (24 h/7 days)
Datamart http://dd.weather.gc.ca/analysis/precip	CaPA–HRDPA, 6 h and 24 h analyses and confidence index (both early and late analyses)	GRIB2 binary files on a 2.5 km polar stereographic grid	Last 32 days	Latest operational version	Operational support by MSC (24 h/7 days)
	CaPA–HRDPA, 6 h and 24 h analyses and confidence index (both early and late analyses)	Watershed averages in shapefile format	Last 32 days	Latest operational version	Operational support by MSC (24 h/7 days)
	CaPA–RDPA, 6 h and 24 h analyses and confidence index (both early and late analyses)	GRIB2 binary files on a 10 km polar stereographic grid	Last 32 days	Latest operational version	Operational support by MSC (24 h/7 days)
CMC’s Collaboration website http://collaboration.cmc.ec.gc.ca/science/outgoing/capa.grib/	CaPA–RDPA, 6 h and 24 h final analyses and confidence index	GRIB2 binary files	April 2011 to present	Archive of operational products	Supported by MSC on a best effort basis
	CaPA–RDPA, 6 h and 24 h final analyses and confidence index	GRIB2 binary files on a 10 km polar stereographic grid	Jan. 2002 to June 2012	CaPA 2.4β8 reanalysis	Supported by MSC on a best effort basis
	CaPA–RDRS, 24 h final analyses and confidence index	GRIB2 binary files on a 15 km polar stereographic grid	Jan. 2010 to Dec. 2014	CaPA–RDRS 1.0 reanalysis	Supported by MSC on a best effort basis
Geomet http://geo.weather.gc.ca/geomet/?service=WMS&request=GetCapabilities	CaPA–RDPA and CaPA–HRDPA, 6 h and 24 h analyses and confidence index (both early and late analyses)	Web mapping service (WMS) compliant with the Open Geospatial Consortium standard	Last 32 days	Latest operational version	Operational support by MSC (24 h/7 days)
MRCC’s Binational Precipitation Tool http://mrcc.isws.illinois.edu	CaPA–RDPA 24 h analyses	Interactive maps and text files	Jan. 2002 to present	CaPA 2.4β8 until May 2011 and operational data after	Supported by MRCC on a best effort basis

Mahfouf, J.-F., Brasnett, B., & Gagnon, S. (2007). A Canadian Precipitation Analysis (CaPA) project: Description and preliminary results. Atmosphere-Ocean, 45(1), 1–17. doi: 10.3137/ao.v450101

 Web of Science ®Google Scholar

Evans, A. M. (2013). Investigation of enhancements to two fundamental components of the statistical interpolation method used by the Canadian Precipitation Analysis (CaPA) (Master’s thesis). University of Manitoba, Winnipeg, Manitoba.

 Google Scholar

Carrera, M., Bélair, S., Fortin, V., Bilodeau, B., Charpentier, D., & Doré, I. (2010). Evaluation of snowpack simulations over the Canadian Rockies with an experimental hydrometeorological modeling system. Journal of Hydrometeorology, 11(5), 1123–1140. doi: 10.1175/2010JHM1274.1

 Web of Science ®Google Scholar

Lespinas, F., Fortin, V., Roy, G., Rasmussen, P., & Stadnyk, T. (2015). Performance evaluation of the Canadian Precipitation Analysis (CaPA). Journal of Hydrometeorology, 16(5), 2045–2064. doi: 10.1175/JHM-D-14-0191.1

 Web of Science ®Google Scholar

Fortin, V., Roy, G., Donaldson, N., & Mahidjiba, A. (2015). Assimilation of radar quantitative precipitation estimations in the Canadian Precipitation Analysis (CaPA). Journal of Hydrology, 531, 296–307. doi: 10.1016/j.jhydrol.2015.08.003

 Web of Science ®Google Scholar

Boluwade, A., Zhao, K.-Y., Stadnyk, T., & Rasmussen, P. (2018). Towards validation of the Canadian Precipitation Analysis (CaPA) for hydrologic modeling applications in the Canadian Prairies. Journal of Hydrology, 556, 1244–1255. doi: 10.1016/j.jhydrol.2017.05.059

 Web of Science ®Google Scholar

Bajamgnigni Gbambie, A. S., Poulin, A., Boucher, M.-A., & Arsenault, R. (2017). Added value of alternative information in interpolated precipitation datasets for hydrology. Journal of Hydrometeorology, 18(1), 247–264. doi: 10.1175/JHM-D-16-0032.1

 Web of Science ®Google Scholar

Hanes, C. C., Jain, P., Flannigan, M. D., Fortin, V., & Roy, G. (2017). Evaluation of the Canadian Precipitation Analysis (CaPA) to improve forest fire danger rating. International Journal of Wildland Fire, 26(6), 509–522. doi: 10.1071/WF16170

 Web of Science ®Google Scholar

Wong, J. S., Razavi, S., Bonsal, B. R., Wheater, H. S., & Asong, Z. E. (2017). Inter-comparison of daily precipitation products for large-scale hydro-climatic applications over Canada. Hydrology and Earth System Sciences, 21(4), 2163–2185. doi: 10.5194/hess-21-2163-2017

 Web of Science ®Google Scholar

Eum, H.-I., Dibike, Y., Prowse, T., & Bonsal, B. R. (2014). Inter-comparison of high-resolution gridded climate data sets and their implication on hydrological model simulation over the Athabasca Watershed, Canada. Hydrological Processes, 28(14), 4250–4271. doi: 10.1002/hyp.10236

 Web of Science ®Google Scholar