Synopsis

This analysis references the U.S. Geological Survey (USGS) earthquake database to analyse the 2020 earthquake in Puerto Rico, particularly the aftershocks and compare it to other strong earthquakes in the region. We also compare Puerto Rico’s seismic activity to other active areas with the goal of gaining a better understanding of how this natural phenomenon might impact the lives of the island’s residents in the future.

How long will the aftershocks last? It is impossible to predict with great accuracy how long aftershocks will be felt, but experts say it can range from a couple of months up to a year. How long have aftershocks lasted after other major earthquakes? Puerto Rico’s 2020 earthquake (6.4 in magnitude) was compared to Haiti’s 2010 (7.0 magnitude) and Mexico’s 2017 (8.2 magnitude) devastating earthquakes.

Will earthquakes be part of Puerto Rico’s new normal? Puerto Rico has always had frequent seismic activity, usually felt like mild tremors. How does Puerto Rico’s sesimic activity compare to that of Chile and Fiji?

When are earthquakes felt?

The way earthquakes are felt are a combination of magnitude, depth and type of movement. Shallow earthquakes tend to cause more damage, while deeper earthquakes may be felt farther along the region.3 We will consider both magnitude and depth in the analysis below. Earthquakes can be felt if they are of 3.0 or higher, though these tend to be very mild and you can barely walk with a 7.0 earthquake.2 For the purpose of this analysis, we only consider events 3.0 and higher.

Data Processing

Libraries

We will use the following libraries to explore and analyze the data:

library(tidyr)
library(dplyr)
library(ggplot2)
library(viridis)
library(lubridate)
library(gridExtra)
library(geosphere)
library(lubridate)
library(ggmap)
library(maps)
library(mapdata)
library(sf)

Downloading Data

Data is downloaded from the USGS website. The website has a minimum of entries that can be downloaded in a single CSV, so data was downloaded in the following chunks:

  1. Puerto Rico and the Caribbean- Downloaded all earthquakes 3.0 and above for the Caribbean since 1975
  2. Mexico- Downloaded all earthquakes 3.0 and above since 2000
  3. Chile- Downloaded all earthquakes 3.0 and above since 2000
  4. Fiji- Downloaded all earthquakes 3.0 and since 2000
##Load USGS Data for Caribbean, Mexico, Chile and Fiji into data frames
dirName <- "data"
caribbeanFileName <- "USGS_Caribbean_historic.csv"
meXicoFileName <- "USGS_Mexico_since_1975.csv"
fijiFileName <- "USGS_Fiji_since_2000.csv"
chileFileName <- "USGS_Chile_since_2000.csv"
ruptureLengthsFileName <- "RuptureLengths.csv"


dirPath <- paste0("./",dirName)
caribbeanFilePath <- paste0(dirPath,"/",caribbeanFileName)
meXicoFilePath <- paste0(dirPath,"/",meXicoFileName)
chileFilePath <- paste0(dirPath,"/",chileFileName)
fijiFilePath <- paste0(dirPath,"/",fijiFileName)
ruptureLengthsFilePath <- paste0(dirPath,"/",ruptureLengthsFileName)
if(!file.exists(dirName)){
  dir.create(dirName)
}
caribbean_usgs_data <- read.csv(caribbeanFilePath)
mexico_usgs_data <- read.csv(meXicoFilePath)
chile_usgs_data <- read.csv(chileFilePath)
fiji_usgs_data <- read.csv(fijiFilePath)
rupture_lengths_data <-  read.csv(ruptureLengthsFilePath)

##merge both data into a single frame
usgs_data <- rbind(caribbean_usgs_data, mexico_usgs_data, chile_usgs_data, fiji_usgs_data)

Data Dictionary

The Data Dictionary for the data can be found in: https://earthquake.usgs.gov/earthquakes/feed/v1.0/csv.php#data

Examining Data

The merged usgs dataset has 55,757 observations of 31 variables including time of event, latitude, longitude, depth and magnitude of the earthquake.

PreProcessing Data

The following transformations were needed in order to complete the required data analysis:

  1. Extract country from “place” variable
  2. Rename untidy fields
  3. Add new fields to tag earthquake category and focal depth
  4. Identify the mainshock (earthquake with highest magnitude) in each country
  5. Identify foreshocks and aftershocks for each mainshock

1. Extract Country and rename fields

########################
##Data Transformations
########################

##Extract country from place. There are two different formats depending on the time of event: 
## a) e.g. Puerto Rico region
## b) e.g. 11km SSE of Guanica, Puerto Rico 
usgs_data <- usgs_data %>% mutate(country = sapply(strsplit(as.character(place),", "), function(x) x[2]))
usgs_data <- usgs_data  %>% mutate(country = ifelse(is.na(country),sapply(strsplit(as.character(place)," region"), function(x) x[1]),country))

2. Rename fields

################
##Rename data 
################

##usgs_data <- usgs_data %>% rename(event_time = "ï..time")
rupture_lengths_data <- rupture_lengths_data %>% rename(country = ï..country)

3. Add new fields to tag earthquake category and focal depth

#####################################
##Tag magnitude & depth categories
#####################################

##magnitude: 
## Source: https://www.britannica.com/science/earthquake-geology/Earthquake-magnitude
usgs_data <- usgs_data  %>% mutate(category = case_when
                                   (mag < 3.0 ~ "2- micro", 
                                     mag >= 3.0 & mag < 4.0 ~ "3- minor", 
                                     mag >= 4.0 & mag < 5.0 ~ "4- light",
                                     mag >= 5.0 & mag < 6.0 ~ "5- moderate",
                                     mag >= 6.0 & mag < 7.0 ~ "6- strong", 
                                     mag >= 7.0 & mag < 8.0 ~ "7- major",
                                     mag >= 8.0 ~ "8+- great"))


##depth
##Source: https://www.usgs.gov/natural-hazards/earthquake-hazards/science/determining-depth-earthquake?qt-science_center_objects=0#qt-science_center_objects

usgs_data <- usgs_data  %>% mutate(focal_depth = case_when
                                   (depth < 30 ~ "Very Shallow", 
                                    depth >= 30 & depth < 70 ~ "shallow", 
                                    depth >= 70 & depth < 300 ~ "intermediate", 
                                    depth >= 300 ~ "deep"))

4. Identify the mainshock (earthquake with highest magnitude) in each country

Note: There were two earthquakes in Mexico and we separated the data for each earthquake.

################################################
##Identify largest earthquake for each country
###############################################

##Get max quake epicanter per country
max_mag_per_ctry <- usgs_data %>% 
  group_by(country) %>% 
  summarise(mainshock_magnitude = max(mag), 
            rank = which.max(mag), 
            latitude_epicenter = latitude[rank],  
            longitude_epicenter = longitude[rank], 
            mainshock_event_time = time[rank] )

##Relate epicenter to rupture length
max_mag_per_ctry <- merge(max_mag_per_ctry, rupture_lengths_data, by = "country", all=TRUE)

##Find all related foreshocks and aftershocks
usgs_data <-  merge(usgs_data, max_mag_per_ctry, by = "country", all= TRUE)

##Filter for Haiti, Mexico and Puerto Rico which will be used for the aftershock analysis
large_earthquake_data <- usgs_data %>% 
  filter(country %in% c("Haiti", "Mexico", "Mexico (Puebla)", "Puerto Rico")) %>% 
  select(id, country, time, mag, depth, longitude, latitude, place, mainshock_magnitude,longitude_epicenter, latitude_epicenter, mainshock_event_time, rupture_length, category)

##Add distance from epicenter
large_earthquake_data <- large_earthquake_data %>% mutate(epicenter_distance = distHaversine(large_earthquake_data[,6:7], large_earthquake_data[,10:11]))
##Add time difference from max earthquake
large_earthquake_data <- large_earthquake_data %>% 
  mutate(mainshock_event_date = as.POSIXct(mainshock_event_time, format="%Y-%m-%dT%H:%M:%S"), this_event_date = as.POSIXct(time, format="%Y-%m-%dT%H:%M:%S")) %>% 
  mutate( mainshock_time_diff_days = difftime(this_event_date,mainshock_event_date, units = "days"), 
          mainshock_time_diff_weeks = difftime(this_event_date,mainshock_event_date, units = "weeks"),
          year = year(this_event_date))

5. Identify foreshocks and aftershocks for each mainshock

Estimated Rupture lengths for each major earthquake were obtained from different websources (see References section) and were used to be able to identify foreshocks and aftershocks. This data may not be exact and does not represent oficiall USGS reports. USGS Data did not include rupture length.

#########################################
##Tag foreshocks and aftershocks
#########################################

##earthquake foreshocks and aftershocks occur within 2 rupture lengths from the epicenter
##Haversine is in meters and data is in kilometers
large_earthquake_data <- large_earthquake_data %>% 
  mutate(near_epicenter_location = epicenter_distance <= (2*1000*rupture_length))


large_earthquake_data <- large_earthquake_data %>% mutate(aftershock = (near_epicenter_location == TRUE & mainshock_time_diff_days >= 0 & mainshock_time_diff_days  < 365))
large_earthquake_data <- large_earthquake_data %>% mutate(foreshock = (near_epicenter_location == TRUE & mainshock_time_diff_days < 0 & mainshock_time_diff_days  > -30))

Reviewing Puerto Rico’s Seismic Activity

Puerto Rico has very active seismic activity, since it lies in a dynamic plate-boundary zone between two tectonic plates: the North American plate and the northeast corner of the Caribbean plate. The majority of the earthquakes are minor, meaning that they can be felt as tremors in some parts of the island. It had been 100 years since a strong earthquake had been felt in Puerto Rico.

A summary of Puerto Rico’s seismic activity since 2016 is depicted in the graph below. Although there is typically more activity in the north, the strong earthquake ocurred in the southern region.

##Get All Seismic activity for Puerto Rico, Chile and Fiji
seismic_activity <- usgs_data %>% 
  mutate( this_event_date = as.POSIXct(time, format="%Y-%m-%dT%H:%M:%S"), year = year(this_event_date) ) %>%
  filter(country %in% c("Fiji", "Chile", "Puerto Rico")) %>% 
  select(id, country, time, mag, depth, longitude, latitude, place, this_event_date, year, category, focal_depth)

##Map only Puerto Rico seismic activity
pr_map <- map_data("world", region = "Puerto Rico")
pr_seismic_activity <- seismic_activity %>%
  filter(country == "Puerto Rico" & year > 2016)


pr_recent_earthquakes_plot <- 
  ggplot() +
  geom_map(data = pr_map, map = pr_map, aes(x = long, y=lat, group=group, map_id=region), fill="white", colour="#7f7f7f", size=0.5) +
  coord_fixed(1) +
  geom_point(data = pr_seismic_activity[pr_seismic_activity$category == "3- minor",],
             aes(x = longitude,
                 y = latitude,
                 color = category,
                 size = mag)) +
  geom_point(data = pr_seismic_activity[pr_seismic_activity$category == "4- light",],
             aes(x = longitude,
                 y = latitude,
                 color = category,
                 size = mag)) +
  geom_point(data = pr_seismic_activity[pr_seismic_activity$category == "5- moderate",],
             aes(x = longitude,
                 y = latitude,
                 color = category,
                 size = mag)) +
  geom_point(data = pr_seismic_activity[pr_seismic_activity$category == "6- strong",],
             aes(x = longitude,
                 y = latitude,
                 color = category,
                 size = mag)) +
 scale_color_manual(values=c("azure3", "deepskyblue3", "deepskyblue4", "blue4")) +
   labs(x="Longitude", y="Latitude",
       title = "Puerto Rico Earthquakes Since 2017",
       caption = "Data source: USGS")

pr_recent_earthquakes_plot

Aftershocks After Earthquakes

Many people in Puerto Rico were not aware or prepared to endure the frequent aftershocks following the main quake. Aftershocks caused the residents a lot of anxiety and some even went to great lengths and slept in their front lawns or cars because they felt safer than in their homes.

How long should we expect to experience aftershocks? According to a USGS study9 aftershocks may be felt for up to a decade. While it is impossible to predict, We can examine the aftershock data for the devastating earthquakes in Haiti (2010) and Mexico (2017) and see what their people had to go through.

##########################################################################################
##Separate USGS Data for Puerto Rico 2020 quake, Mexico 2017 quake, and Haiti 2010 quake
#########################################################################################
large_earthquake_data <-  large_earthquake_data %>% 
  filter(aftershock == TRUE | foreshock == TRUE) %>% 
  arrange(this_event_date)

country.labs <- c("Haiti 2010", "Mexico (Chiapas) 2017", "Mexico (Puebla) 2017", "Puerto Rico 2020")
names(country.labs) <- c("Haiti", "Mexico", "Mexico (Puebla)", "Puerto Rico")

large_quake_plot <- ggplot(large_earthquake_data, aes(x=mainshock_time_diff_days, y=mag, color = country)) +
  geom_line() +
  geom_smooth (method = "lm", se = FALSE, color = "azure4") +
  scale_color_discrete(name="Country",
                      breaks=c("Haiti", "Mexico", "Mexico (Puebla)", "Puerto Rico"),
                      labels=country.labs) +
  facet_wrap(~ country, ncol = 1, scales = "fixed", labeller = labeller(country = country.labs)) +
  labs(x="Days After Mainshock", y="Magnitude", title = "Strong Earthquakes",
       subtitle = "Foreshocks and aftershocks with a magnitude > 3.0",
       caption = "Earthquake Data source: USGS. Foreshocks and Aftershocks are estimated.") +
  theme(plot.title = element_text(hjust = 0.5), 
        plot.subtitle = element_text(color = "azure4",hjust = 0.5),
        plot.caption = element_text(color="azure4")) 

large_quake_plot

Haiti’s 2010 earthquake was of 7.0 magnitude and it caused major devastation. It killed 230,000 people, injured 300,000 and damaged over 400,00 structures. Haiti did not have a lot of seismic activity following the mainshock, but the magnitude caused a longlating impact, and Haiti has not yet fully recovered.

Mexico suffered two major earthquakes in 2017 just two weeks part, one 8.2 magnitude near Chiapas, the other 7.1 near Puebla. Thousands of structures were damaged, including schools, and around 300 people were killed. Although the Puebla quake caused much more devastation, the Chiapas quake had more frequent seismic activity and moderate 5.0 aftershocks ocurred monthly almost a year after the main earthquake.

Puerto Rico’s 2020 earthquake was of 6.4 magnitude and caused significant damage to structures in the southern region, including collapsing a school in Guanica. In the three months following the earthquake, the seismic activity was even more frequent than in Chiapas. We will have to wait and see if the trend is similar to Chiapa’s or if the frequency decreases like it did in Haiti and Puebla.

Aftershock occurrences per month.

###########################################################
##Plot aftershock magnitudes and occurrences per month
###########################################################

##months defined as 30 day slot for simplicity sake
large_quake_aftershocks <- large_earthquake_data %>% 
  filter(country %in% c("Mexico","Mexico (Puebla)", "Haiti","Puerto Rico") & (aftershock == TRUE | foreshock == TRUE)) %>%  
  group_by(country,months = round(mainshock_time_diff_days/30)) %>%  
  summarise(occurrences = n(), max_mag = max(mag))

large_quake_aftershock_plot <- ggplot(large_quake_aftershocks, aes(x=months, y=max_mag, size=occurrences, color = country)) +
  geom_point(alpha=0.5) +
  scale_size(range = c(.1, 24), name="Occurrences") +
  labs(x="Months After Mainshock", y="Maximum Magnitude for the Month", 
       title = "Aftershock Occurrences in Strong Earthquakes",
       caption = "Earthquake Data source: USGS. Foreshocks and Aftershocks are estimated.",
       fill = "Country") +
  theme(plot.title = element_text(hjust = 0.5), 
        plot.subtitle = element_text(color = "azure4",hjust = 0.5),
        plot.caption = element_text(color="azure4")) +
  scale_color_discrete(name="Country",
                       breaks=c("Haiti", "Mexico", "Mexico (Puebla)",  "Puerto Rico"),
                       labels=country.labs)

large_quake_aftershock_plot

Regions with Frequent Seismic Activities

According to the USGS, Japan and Indonesia are two of the countries with the more recent earthquakes.1 Fiji has a lot of seismic activity, and it being an island, may be a better comparison to Puerto Rico. Chile also has frequent earthquakes and given the resiliency of their people and their calm reactions to 6.8 quakes, we deemed it appropriate to include it in our analysis. Watching this YouTube Video will show the kind of resiliency we are talking about. https://www.youtube.com/watch?v=aIyf7t_hcZU

Understanding focal depth

Both earthquake depth and magnitude are contributing factors to the strength of shaking in affected regions. The most devastating earthquakes are usually under 30km deep and they need to be at least 3.0 in magnitude to be felt.

The following maps show the sesimic activity and focal depth for Chile, Fiji and Puerto Rico in the last three years.

#################################################################
##Map seismic activity by focal depth for each region
################################################################

##Chile
chile_map <- map_data("world", region = "Chile")
chile_recent_seismic_activity <- seismic_activity %>%
  filter(country == "Chile" & year > 2016)


chile_recent_quakes_bydepth <- 
    ggplot() +
    geom_map(data = chile_map, map = chile_map, aes(x = long, y=lat, group=group, map_id=region), fill="white", colour="#7f7f7f", size=0.5) +
    ##coord_fixed(1.3) +
    coord_sf(xlim = c(-81, -65), ylim = c(-58,-20), expand = FALSE) +
    geom_point(data = chile_recent_seismic_activity,
             aes(x = longitude,
                 y = latitude,
                 color = focal_depth)) +
 ## scale_color_manual(values=c("bisque", "coral", "brown3", "darkred")) +
  scale_color_manual(values=c("grey50", "steelblue3", "mediumseagreen")) +
  labs(x="Longitude", y="Latitude", title = "Chile") +
  theme(plot.title = element_text(hjust = 0.5), 
        plot.subtitle = element_text(color = "azure4",hjust = 0.5),
        plot.caption = element_text(color="azure4"))



##Fiji
world_map <- map_data("world")
fiji_recent_seismic_activity <- seismic_activity %>%
  filter(country == "Fiji" & year > 2016)

##the map region did not work so I had to use coord_sf to map by coordinates
fiji_recent_quakes_bydepth <-
  ggplot() +
  geom_map(data = world_map, map = world_map, aes(x = long, y=lat, group=group, map_id=region), fill="white", colour="#7f7f7f", size=0.5) +
 ## coord_sf(xlim = c(170, 183), ylim = c(-25,-10), expand = FALSE) +
  coord_sf(xlim = c(174, 182), ylim = c(-22,-13), expand = FALSE) +
  geom_point(data = fiji_recent_seismic_activity,
             aes(x = longitude,
                 y = latitude,
                 color = focal_depth)) +
  ##scale_color_manual(values=c("azure3", "darkseagreen3", "springgreen3", "darkgreen")) +
  scale_color_manual(values=c("black", "grey50", "steelblue3", "mediumseagreen")) +
  labs(x="Longitude", y="Latitude", title = "Fiji") +
  theme(plot.title = element_text(hjust = 0.5), 
        plot.subtitle = element_text(color = "azure4",hjust = 0.5),
        plot.caption = element_text(color="azure4"))


##Puerto Rico

pr_recent_quakes_bydepth <-
  ggplot() +
  geom_map(data = pr_map, map = pr_map, aes(x = long, y=lat, group=group, map_id=region), fill="white", colour="#7f7f7f", size=0.5) +
  coord_fixed(1) +
  geom_point(data = pr_seismic_activity[pr_seismic_activity$focal_depth=="Very Shallow",],
             aes(x = longitude,
                 y = latitude,
                 color = focal_depth)) +
  
  geom_point(data = pr_seismic_activity[pr_seismic_activity$focal_depth!="Very Shallow",],
             aes(x = longitude,
                 y = latitude,
                 color = focal_depth)) +
  #scale_color_manual(values=c("steelblue3", "deepskyblue3", "deepskyblue4", "blue4")) +
  scale_color_manual(values=c("grey50", "steelblue3", "mediumseagreen")) +
  labs(x="Longitude", y="Latitude",
       title = "Puerto Rico") +
  theme(plot.title = element_text(hjust = 0.5), 
        plot.subtitle = element_text(color = "azure4",hjust = 0.5),
        plot.caption = element_text(color="azure4"))



#extract legend
#https://github.com/hadley/ggplot2/wiki/Share-a-legend-between-two-ggplot2-graphs
g_legend<-function(a.gplot){
  tmp <- ggplot_gtable(ggplot_build(a.gplot))
  leg <- which(sapply(tmp$grobs, function(x) x$name) == "guide-box")
  legend <- tmp$grobs[[leg]]
  return(legend)}

my_legend<-g_legend(fiji_recent_quakes_bydepth)

grid.arrange(  pr_recent_quakes_bydepth + theme(legend.position="none"),
               my_legend,
               ncol=2,
               widths= c(3/4, 1/4))

grid.arrange( chile_recent_quakes_bydepth + theme(legend.position="none"),
              fiji_recent_quakes_bydepth + theme(legend.position="none"),
              ncol=2,
              widths=c(1/3,2/3))

Recent Seismic Activity

Is Puerto Rico’s seismic activity comparable to Chile’s and Fiji’s? While Puerto Rico has frequent activity, both Chile and Fiji experience earthquakes that are stronger in magnitude, upwards of 5.0. Puerto Rico’s earthquakes are generally below 4.0 with the exception of the months before the January 2020 earthquake, as can be seen in the graph below.

recent_seismic_activity <- seismic_activity %>%
  filter(year > 2016 & year < 2020)  %>%
  group_by(country, year,month = month(this_event_date, abbr = TRUE, label = TRUE)) %>%
  summarize(occurrences = n(), max_mag = max(mag))

recent_seismic_activity_plot <-
  ggplot(recent_seismic_activity, aes(x=month, y=max_mag, size=occurrences, color = country)) +
  geom_point(alpha=0.5) +
  scale_size(range = c(.1, 24), name="Occurrences") +
  facet_wrap(~ year, ncol = 1, scales = "fixed") +
  labs(x="Month", y="Maximum Magnitude for the Month", title = "All  Seismic Activity in Chile, Fiji and Puerto Rico",
       subtitle = "(All regions, All Depths)",
       caption = "Data source: USGS") +
  theme(plot.title = element_text(hjust = 0.5), 
        plot.subtitle = element_text(color = "azure4",hjust = 0.5),
        plot.caption = element_text(color="azure4"))

recent_seismic_activity_plot

Results

Puerto Rico’s aftershocks are more frequent than those of Haiti 2010 earthquakes and Mexico 2017 earthquakes. As explained by USGS9, aftershocks may continue for months, years, and up to a decade.

Puerto Rico’s seismic activity is mild compared to Chile’s and Fiji’s. If other tectonic plates release energy in the next few years, we may be joining those countries and hopefully we will have more resilient structures and a seismic culture that can withstand the natural phenomena.

References

  1. Which Country Has The Most Earthquakes?. USGS. https://www.usgs.gov/faqs/which-country-has-most-earthquakes?qt-news_science_products=0#qt-news_science_products

  2. Bolt, Bruce. Earthqake Magnitude. January 29,2020. Encyclopedia Britannica. https://www.britannica.com/science/earthquake-geology/Shallow-intermediate-and-deep-foci

  3. Chang, Alicia. Why shallow earthquakes like the one in Italy tend to cause more damage than deep ones. The Associated Press. August 24, 2016. https://www.businessinsider.com/ap-ap-explains-difference-between-shallow-deep-earthquakes-2016-8

  4. William Spence, Stuart A. Sipkin, and George L. Choy. Determining the Depth of an Earthquake. Earthquakes and Volcanoes. Volume 21, Number 1, 1989. https://www.usgs.gov/natural-hazards/earthquake-hazards/science/determining-depth-earthquake?qt-science_center_objects=0#qt-science_center_objects

  5. Amadeo, Kimberly. Haiti Earthquake Facts, Its Damage and the Effects on the Economy.. The Balance. December 27, 2019. https://www.thebalance.com/haiti-earthquake-facts-damage-effects-on-economy-3305660.

  6. World Vision Staff. 2017 Mexico earthquakes: Facts, FAQs, and how to help. September 21, 2017. https://www.worldvision.org/disaster-relief-news-stories/2017-mexico-earthquakes-facts

  7. Wikipedia. 2010 Haiti Earthquake. https://en.wikipedia.org/wiki/2010_Haiti_earthquake. Accessed March 8, 2020.

  8. Ye, Lingling, Thorne Lai, Bai F., Cheung K.F, Kanamori H. The 2017 Mw 8.2 Chiapas, Mexico, Earthquake: Energetic Slab Detachment. Geophysical Research Letters, 44. https://doi.org/10.1002/2017GL076085

  9. van der Elst, N.J., Hardebeck, J.L., and Michael, A.J., 2020, Potential duration of aftershocks of the 2020 southwestern Puerto Rico earthquake: U.S. Geological Survey Open-File Report 2020–1009, 5 p., https://doi.org/10.3133/ofr20201009.