This is a daily updated tracker website providing time-varying estimated reproduction numbers for COVID-19 in California, based on both daily case counts and wastewater readings.
We estimate transmission rates of COVID–19 using reproduction numbers. A time-varying reproduction number, commonly written R, is the average number of cases infected by a given case over the course of that individual’s disease progression. Of particular interest is the question of whether transmission is supercritical, meaning that R > 1, in which case the epidemic can increase in size, or is subcritical, meaning that R < 1, in which case it will fade out. To eliminate a disease locally, it is not necessary to reduce R to zero, only to reduce it below one for a sustained period.
We use raw data from three sources: new cases by county from California Health and Human Services, (“Cases”, below); wastewater readings from the Cal-SuWers network (“Wastewater (C)”), WastewaterScan (“Wastewater (W)”), and BioBot Analytics (“Wastewater (B)”).
For more details see the Methods section below.
Daily new cases by county are downloaded daily from the California Health and Human Services Open Data Portal.
Wastewater readings are downloaded from the data.ca.gov
Open Data Portal and from WastewaterScan each
day. Counts of the n
and s
gene targets are
used, normalized to counts of the pepper mottle virus. Counts for which
the PPMoV control is not present are excluded. Each count is used in all
counties included in its wastewater collection region. The wastewater
estimates published by BioBot
Analytics are included as well, because they include some additional
locations. Exceptionally high wastewater counts are indicated by
^
signs at the top of each plot.
We apply kernel density smoothing with a Gaussian kernel with bandwidth of 14 days to both wastewater and case counts, to de-emphasize day-to-day fluctuations and pick out longer-term changes.
The Wallinga-Teunis technique is used to estimate reproduction numbers, corrected for right censoring of the case detection data, and a small number of the earliest days are trimmed from the estimates to avoid artifacts due to left censoring. Periods when case numbers are less than 2 per day in a county are excluded from estimation from case numbers. The generation interval distribution used for estimation is that estimated here. The generation time estimate for the Delta variant is used for dates before Dec. 24, 2021, and for Omicron is used thereafter. Thus the estimates may have a small bias in the case of variants with other transmission intervals.
Lee Worden, F.I. Proctor Foundation, [email protected]
Alex Y. Ge, UCSF School of Medicine
Micaela Neus, Woodlamp Technology
Jennifer C. Kwan, Woodlamp Technology
Nathan Murthy, Tesla
Jianda Monique
Eugene T. Richardson, Department of Medicine, Brigham and Women’s
Hospital, Boston, MA, Department of Global Health and Social Medicine,
Harvard University, Boston, MA
Rae Wannier, F.I. Proctor Foundation, UCSF Dept. of Epidemiology and
Biostatistics
Travis Porco, F.I. Proctor Foundation, UCSF Dept. of Epidemiology and
Biostatistics
This is not an official UCSF website. The opinions or statements expressed herein should not be taken as a position of or endorsement by the University of California, San Francisco.
Last updated : 2024-11-20 01:16:17 PST