weatherOz Use Case

Introduction

Summer pulses (e.g., mungbeans) are a high value crop for Queensland grain growers. However, temperatures above 40 °C, particularly during flowering, can have a severe detrimental effect on crop yields (Kaushal et al. 2016). A key strategy that plants use to minimise the impact of heat stress is evaporative cooling. Ensuring crops are well watered prior to a heat stress event (so to facilitate high rates of evaporative cooling) is a potential strategy to ensure that irrigated crops avoid heat stress. However, this requires farmers to recognise a heat stress event is coming and commence irrigating with enough time to water the entire crop prior to the event occurring.

The {WINS} for pulses project provides warnings when heat stress events are likely to occur so that farmers can irrigate their crops prior to the event to minimise any resulting yield loss. It does this by using the get_precis_forecast() function from {weatherOz} along with functionality from {mailR} (Premraj 2021).

Materials and Methods

install.packages("weatherOz")
install.packages("mailR")
library(dplyr)  # for filter()
library(weatherOz)
library(mailR)

For this to work we require a list of alert subscribers. Here we create a dummy set of subscribers, locations, and email addresses.

subscribers_list <-
  data.frame(cbind(
    c(1, 2, 3),
    c("Joe", "John", "Jayne"),
    c("Blogs", "Doe", "Doe"),
    c("Dalby", "Toowoomba", "Warwick"),
    c(
      "[email protected]",
      "[email protected]",
      "[email protected]"
    )
  ))
colnames(subscribers_list) <-
  c("Entry", "Name", "Surname", "Location", "email")
head(subscribers_list)

##   Entry  Name Surname  Location                 email
## 1     1   Joe   Blogs     Dalby   [email protected]
## 2     2  John     Doe Toowoomba [email protected]
## 3     3 Jayne     Doe   Warwick     [email protected]

We also need to specify our email address and password. For this project we are using a gmail account to send the email alerts.

our_email <- "yyyy***[email protected]"
our_password <- "*password*"

Finally we need to set a forecast temperature threshold that will trigger an alert to be sent.

threshold_temp <- 40

We then use {weatherOz} to access the latest précis forecast for Queensland.

QLD_forecast <- get_precis_forecast(state = "QLD")

Now we use a for() loop that searches through the Queensland précis forecast data frame and subsets it based on location each subscriber has nominated and the temperature threshold we have set. As part of the for() loop there is an if() statement that uses the {mailR} package to send an email to the subscribers if the forecast maximum temperatures are greater than or equal to the threshold. This email informs the recipient of the dates that high temperatures are expected.

QLD_hotdates <-
  QLD_forecast %>%
  filter(maximum_temperature >= threshold_temp)

for (x in seq_len(nrow(subscribers_list))) {
  subscriber_location <- subscribers_list[["Location"]][x]
  if (subscriber_location %in% QLD_hotdates[["town"]]) {
    hot_dates <-
      paste(gsub("00:00:00", "", QLD_hotdates$start_time_local),
            collapse = ", ")

    body_text <-
      paste(
        "\nHello ",
        as.character(subscribers_list$Name[x]),
        ".\n",
        "\nYour mungbean crops at ",
        subscriber_location,
        "\nare forecast to be exposed to heat stress on the\n",
        "\nfollowing dates: ",
        hot_dates,
        ".\n",
        "\nConsider irrigating your crops beforehand to\n"
        "\nfacilitate transpirational cooling.\n",
        "\nFrom the WINS team\n"
      )
    recipient <-
      as.character(subscribers_list$email[x])
    send.mail(
      from = our_email,
      to = recipient,
      subject = "Mungbean Heat Stress Warning",
      body = body_text,
      smtp = list(
        host.name = "smtp.gmail.com",
        port = 465,
        user.name = our_email,
        passwd = our_password,
        ssl = TRUE
      ),
      authenticate = TRUE,
      send = TRUE
    )
  }
}

Summary

The process is automated to run on a cron job such that the précis forecast is automatically downloaded and thresholds are calculated on a daily basis with email alerts being sent automatically.

While mungbean heat stress is illustrated here, other thresholds can be implemented, e.g., frost events for sugarcane harvest using this same functionality.