In markwh/rcmodel: Calibrate and Validate Rating Curves for Riverine Load Estimation

knitr::opts_chunk$set(echo = TRUE)

Making some more wqp functions.

1. Map stations

data("aluminumData")
library(leaflet)
library(markstats)

names(attributes(aluminumData))

sites <- attr(aluminumData, "siteInfo")


str(sites, 1)

l1 <- leaflet() %>% 
  addTiles() %>% 
  addRectangles(lng1 = -80, lat1 = 38, lng2 = -77, lat2 = 40)

l1 %>% 
  addCircleMarkers(lng = ~dec_lon_va, lat = ~dec_lat_va, data = sites)

facsum(sites$MonitoringLocationTypeName)


wqp_mapStations <- function(wqpData)

Chasing down a problem with wqp_checkUnits. data.frame index replacement is failing for reasons I don't understand.

# modified output so no3Data is now a list of intermediate objects created within wqp_checkUnits function.

names(no3Data)

no3Data <- nitrateData %>% 
  wqp_checkClasses() %>% 
  wqp_checkActivity() %>% 
  wqp_checkFraction() %>% 
  wqp_checkUnits(convertTo = c("mg/l", "mg/l as N", "meq/l", "mg/kg"),
                 inconvertibles = "omit")

foo <- no3Data$out
dim(foo)

with(no3Data, foo[badrows, colind] <- badunits) # same error

sum(no3Data$badrows)

dim(foo)
sum(no3Data$badrows)
summary(no3Data$badrows)
no3Data$colind
with(no3Data, length(foo[badrows, colind]))

Meanwhile the same operation appears to work with mtcars

bar <- mtcars

logrows <- sample(c(TRUE, FALSE), 32, replace = TRUE)
bar[logrows, 5]
bar[logrows, 5] <- 1:sum(logrows)
bar

Try attaching instead

attach(no3Data)

sum(badrows)
foo[which(badrows), colind] <- 1:sum(badrows)

Using numeric indices instead of logical (i.e. using which) appears to work.

WQP is now down for maintenance. What does the url call return?

foo <- httr::GET("http://www.waterqualitydata.us/Result/search?bBox=-80%2C38%2C-77%2C40&siteType=Stream&characteristicName=Aluminum&sampleMedia=Water&sorted=no&mimeType=tsv")

httr::content(foo, as = "parsed")

Convert to rcData

no3Simple <- wqp_simplifyConc(no3Data)
qSimple <- wqp_simplifyFlow(qData)

intersect(names(no3Simple), names(qSimple))

rawdat1 <- full_join(no3Simple, qSimple, by = c("Date", "datetime", "station"))
summary(rawdat1)

What to do when wq data is reported at higher time resolution than flow data?

• use daily average?
• average wq observations for that day
• Use temporally nearest flow observation
• omit these data

Scoping of duplicate rows

compareDuplicateFlow <- function(simpleData, origData, which = NULL) {

  dupinds <- which(duplicated(simpleData))

  if(is.null(which))
    which <- 1:length(dupinds)
  dups1 <- simpleData[dupinds[which], ] %>% 
    split(f = rownames(.))

  isSame <- function(data, row) {
    data %>% 
      filter(ActivityStartDate == row[["Date"]],
           MonitoringLocationIdentifier == row[["station"]],
           ResultMeasureValue == row[["flow"]],
           ResultStatusIdentifier == row[["flow.flag"]],
           ActivityStartDateTime == row[["datetime"]]) %>% 
      mutate(difRows = paste(whichHasDifs(.), collapse = "; "))
  }

  out <- lapply(dups1, isSame, data = origData) %>% 
    bind_rows()
  out
}

whichHasDifs <- function(df) {
  difinds <- vapply(df, function(col) length(unique(col)) > 1, logical(1))
  names(df)[difinds]
}

dupFlow <- compareDuplicateFlow(qSimple, qData)
glimpse(dupFlow)
summary(as.factor(dupFlow$difRows)) %>% sort()

dupFlow %>% 
  filter(grepl("DepthHeight", difRows)) %>% 
  glimpse()

dupFlow %>% 
  filter(grepl("Media", difRows)) %>% 
  glimpse()

dupFlow %>% 
  filter(grepl("Comment", difRows)) %>% 
  `[[`("ActivityCommentText")

dupFlow %>% 
  filter(grepl("PCode", difRows)) %>% 
  glimpse()

Flow data shouldn't have depth height measure.

Flow media shouldn't be "other".

Mean daily flow shouldn't have a time that's in the middle of the afternoon.

Also, datetime appears to be incorrectly representing time zones. The time portion of it doesn't equal the "ActivityStartTime.Time" value.

• check this out in the original data.

timetest <- nitrateData %>% 
  filter(!is.na(ActivityStartTime.Time)) %>% 
  select(ActivityStartTime.Time,
         ActivityStartTime.TimeZoneCode,
         ActivityStartDateTime) %>% 
  filter(!is.na(ActivityStartTime.TimeZoneCode))

timetest

timetest$ActivityStartDateTime

So ActivityStartDateTime is always in UTC. Good to know.

What I see resulting from this: No important differences are actually represented.

markwh/rcmodel documentation built on May 21, 2019, 12:26 p.m.