Simulate the Spread of Ascochyta Blight in Chickpea.
README
ascotraceR: An R package resource to simulate the spatiotemporal spread of Ascochyta blight in a chickpea field over a growing season
The goal of of ascotraceR is to develop a weather driven model to simulate the spread of Ascochyta blight disease in a chickpea field over a growing season.
This model is adapted from a model developed by (Diggle et al. 2002) for simulating the spread of anthracnose in a lupin field. The model is run using local weather data. The ascotraceR model simulates the pathogen related processes of conidial production, dispersal, successful deposition and infection on chickpea plants. Host related processes of growth are simulated in terms of development of growing points. The model divides the paddock into 1 square metre cells (observation quadrats/units) and simulates chickpea growth and A. rabiei activities in each cell. Initially, there is one growing point per sown seed when seed are sown. Chickpea growth is then described in terms of increase in the number of growing points. Conidia are dispersed from infested stubble by rain splash or wind driven rain when rainfall threshold is reached. Rainfall threshold refers to the minimum amount of rainfall required to disperse conidia from pycnidia and to provide sufficient duration of moisture for conidia to germinate and penetrate into the host tissues. After penetrating host tissues, conidia produce infected growing points. Infected growing points become sporulating lesions after completion of a latent period. The length of the latent period is a function of temperature, and the number of conidia produced per sporulating growing point depends on the level of resistance of the chickpea cultivar. As the model runs, it keeps a continuous track of non-infected, latent, infected and sporulating growing points (lesions). The ascotraceR’s minimum input requirements are location specific weather data and a list of input variables.
Quick start
ascotraceR is not yet on CRAN. You may install it from GitHub this way.
if (!require("remotes"))
install.packages("remotes")
remotes::install_github("IhsanKhaliq/ascotraceR",
build_vignettes = TRUE
)
Once installed you can simulate disease spread in a chickpea paddock.
Load the library.
library("ascotraceR")
library("data.table")
library("lubridate")
Import the weather data
# weather data
Billa_Billa <- fread(
system.file(
"extdata",
"2020_Billa_Billa_weather_data_ozforecast.csv",
package = "ascotraceR"
)
)
# format time column
Billa_Billa[, local_time := dmy_hm(local_time)]
# specify the station coordinates of the Billa Billa weather station
Billa_Billa[, c("lat", "lon") := .(-28.1011505, 150.3307084)]
head(Billa_Billa)
## day local_time assessment_number mean_daily_temp wind_ km_h ws
## 1: 1 2020-06-04 00:00:00 NA 4.1 0.9 0.25
## 2: 1 2020-06-04 00:15:00 NA 3.9 0.6 0.17
## 3: 1 2020-06-04 00:30:00 NA 3.9 2.0 0.56
## 4: 1 2020-06-04 00:45:00 NA 4.3 1.2 0.33
## 5: 1 2020-06-04 01:00:00 NA 3.7 2.4 0.67
## 6: 1 2020-06-04 01:15:00 NA 3.5 1.3 0.36
## ws_sd wd wd_sd cummulative_rain_since_9am rain_mm wet_hours location
## 1: NA 215 NA 0 0 NA Billa_Billa
## 2: NA 215 NA 0 0 NA Billa_Billa
## 3: NA 215 NA 0 0 NA Billa_Billa
## 4: NA 215 NA 0 0 NA Billa_Billa
## 5: NA 215 NA 0 0 NA Billa_Billa
## 6: NA 215 NA 0 0 NA Billa_Billa
## lat lon
## 1: -28.10115 150.3307
## 2: -28.10115 150.3307
## 3: -28.10115 150.3307
## 4: -28.10115 150.3307
## 5: -28.10115 150.3307
## 6: -28.10115 150.3307
Wrangle weather data
A function, format_weather(), is provided to convert raw weather data into the format appropriate for the model. It is mandatory to use this function to ensure weather data is properly formatted before running the model.
Billa_Billa <- format_weather(
x = Billa_Billa,
POSIXct_time = "local_time",
temp = "mean_daily_temp",
ws = "ws",
wd_sd = "wd_sd",
rain = "rain_mm",
wd = "wd",
station = "location",
time_zone = "Australia/Brisbane",
lon = "lon",
lat = "lat"
)
Simulate Ascochyta blight spread
A function, trace_asco(), is provided to simulate the spread of Ascochyta blight in a chickpea field over a growing season.
# Predict Ascochyta blight spread for the year 2020 at Billa Billa
traced <- trace_asco(
weather = Billa_Billa,
paddock_length = 20,
paddock_width = 20,
initial_infection = "2020-07-17",
sowing_date = "2020-06-04",
harvest_date = "2020-10-27",
time_zone = "Australia/Brisbane",
seeding_rate = 40,
gp_rr = 0.0065,
spores_per_gp_per_wet_hour = 0.6,
latent_period_cdd = 150,
primary_inoculum_intensity = 100,
primary_infection_foci = "centre"
)
Summarise the output
You can easily get summary statistics for the whole paddock over the simulated season and area under the disease progress curve, AUDPC, using summarise_trace().
summarise_trace(traced)
## i_day new_gp susceptible_gp exposed_gp infectious_gp i_date day
## 1: 1 16000.00000 16000.00 0 0 2020-06-04 156
## 2: 2 1108.59107 17108.59 0 0 2020-06-05 157
## 3: 3 1334.03497 18442.63 0 0 2020-06-06 158
## 4: 4 1492.26177 19934.89 0 0 2020-06-07 159
## 5: 5 1793.24969 21728.14 0 0 2020-06-08 160
## ---
## 143: 143 31.35426 1992020.37 3 686 2020-10-24 298
## 144: 144 22.60983 1992042.98 4 686 2020-10-25 299
## 145: 145 20.50306 1992063.48 4 686 2020-10-26 300
## 146: 146 15.78047 1992076.26 1 689 2020-10-27 301
## 147: 147 14.40328 1992090.66 2 689 2020-10-28 302
## cdd cwh cr gp_standard AUDPC
## 1: 0.00000 0 0.0 40.00000 45634.5
## 2: 10.74583 0 0.0 42.77148 45634.5
## 3: 22.84583 0 0.0 46.10657 45634.5
## 4: 35.41042 0 0.0 49.83722 45634.5
## 5: 49.38958 1 0.6 54.32034 45634.5
## ---
## 143: 2133.81645 65 76.2 4980.06672 45634.5
## 144: 2154.64770 73 94.4 4980.12110 45634.5
## 145: 2176.49562 73 94.4 4980.17042 45634.5
## 146: 2195.63104 73 94.4 4980.20748 45634.5
## 147: 2215.57687 74 97.0 4980.24131 45634.5
Reference
Diggle AJ, Salam MU, Thomas GJ, Yang H, O’connell M, Sweetingham M, 2002. AnthracnoseTracer: a spatiotemporal model for simulating the spread of anthracnose in a lupin field. Phytopathology 92, 1110-21.