| Title: | Statistical Modeling of Compound Events |
|---|---|
| Description: | Tools for extracting occurrences, assessing potential driving factors, predicting occurrences, and quantifying impacts of compound events in hydrology and climatology. Please see Hao Zengchao et al. (2019) <doi:10.1088/1748-9326/ab4df5>. |
| Authors: | Zengchao Hao |
| Maintainer: | Zengchao Hao <[email protected]> |
| License: | GPL-3 |
| Version: | 0.3.0 |
| Built: | 2024-11-21 05:04:55 UTC |
| Source: | https://github.com/cran/CompoundEvents |
Tools for extracting occurrences, assessing potential driving factors, predicting occurrences, and quantifying impacts of compound events in hydrology and climatology.
Examples of compound events in hydroclimatology include, but not limited to, compound dry-hot events and compound precipitation and surge (or sea level) events. Take the compound dry and hot event as an example. The function GetDH is used for extracting occurrences based on thresholds of dry and hot indicators. The function DriverLGR is used for assessing potential driving factors of compound events based on logistic regression model. The function PredLGR is used for predicting occurrences of compound events. The function ImpactMG is used for quantifying impacts of compound dry and hot events based on meta-Gaussian model.
Zengchao Hao
Hao, Z., et al.(2013). Changes in concurrent monthly precipitation and temperature extremes. Environ. Res. Lett. 8: 034014.
Hao, Z. et al. (2019). A monitoring and prediction system for compound dry and hot events. Environ. Res. Lett., 14:114034.
Hao, Z. et al. (2019). Statistical prediction of the severity of compound dry-hot events based on ENSO . J. Hydrol., 572: 243-250.
Feng, S. et al. (2019). Probabilistic evaluation of the impact of compound dry-hot events on global maize yields. Sci. Total. Environ., 689: 1228-1234.
Use the logistic regression model to establish relationships between climate indices (e.g., ENSO) and occurrences of compound dry-hot events.
DriverLGR(Y,CI)DriverLGR(Y,CI)
Y |
Occurrence of compound dry-hot events (0-1 binary variable) |
CI |
Climate index as the driving factor of compound events (e.g., ENSO) |
slope parameter and associated p-value
Hao, Z. et al. (2019). A monitoring and prediction system for compound dry and hot events. Environ. Res. Lett., 14:114034.
CI=c(-0.7,-1.2,1.3,0.7,-0.6,1.1,-0.5,0.8,0.5,-0.5,1.6,-1.8,-0.5,-1.4,-0.1,2.2,-0.7,-1.1, 0.6, -1.7) Y=c(0,0,1,1,0,0,0,0,0,0,1,0,1,0,0,1,0,0,0,0 ) res<-DriverLGR(Y,CI)CI=c(-0.7,-1.2,1.3,0.7,-0.6,1.1,-0.5,0.8,0.5,-0.5,1.6,-1.8,-0.5,-1.4,-0.1,2.2,-0.7,-1.1, 0.6, -1.7) Y=c(0,0,1,1,0,0,0,0,0,0,1,0,1,0,0,1,0,0,0,0 ) res<-DriverLGR(Y,CI)
Compute univariate empirical probability
Empdis1(mp)Empdis1(mp)
mp |
monthly precipitation |
The empirical probability
Hao, Z. et al., 2019a. Statistical prediction of the severity of compound dry-hot events based on El Ni??o-Southern Oscillation. J. Hydrol., 572, 243-250.
mp=matrix(rnorm(120,0,1),ncol=1) nd<-Empdis1(mp)mp=matrix(rnorm(120,0,1),ncol=1) nd<-Empdis1(mp)
Compute bivariate empirical probability
Empdis2(mp,mt)Empdis2(mp,mt)
mp |
monthly precipitation |
mt |
monthly temperature |
The bivariate empirical probability
Hao, Z. et al., 2019a. Statistical prediction of the severity of compound dry-hot events based on El Ni??o-Southern Oscillation. J. Hydrol., 572, 243-250.
mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) nd<-Empdis2(mp,mt)mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) nd<-Empdis2(mp,mt)
Extract compound dry-cold occurrences based on thresholds of precipitation and temperature. The binary variable of the dry and cold (DC) event can be obtained.
GetDC(mp,mt,threp,thret)GetDC(mp,mt,threp,thret)
mp |
Precipitation |
mt |
Temperature |
threp |
Threshold of precipitation (e.g., 20th percentile) |
thret |
Threshold of temperature (e.g., 20th percentile) |
The occurrence of compound wet-hot event (0-1 binary variable)
Hao, Z. et al (2013). Changes in concurrent monthly precipitation and temperature extremes. Environ. Res. Lett., 8(3): 034014.
mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) threp=20 thret=20 DC<-GetDC(mp,mt,threp,thret)mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) threp=20 thret=20 DC<-GetDC(mp,mt,threp,thret)
Extract compound dry-hot (DH) occurrences based on thresholds of precipitation and temperature. The binary variable of the DH (or dry-warm) event can be obtained.
GetDH(mp,mt,threp,thret)GetDH(mp,mt,threp,thret)
mp |
Precipitation |
mt |
Temperature |
threp |
Threshold of precipitation (e.g., 20th percentile) |
thret |
Threshold of temperature (e.g., 80th percentile) |
The occurrence of compound dry-hot events (0-1 binary variable)
Hao, Z. et al. (2018). A multivariate approach for statistical assessments of compound extremes. J. Hydrol., 565: 87-94.
Hao, Z. et al. (2019). A monitoring and prediction system for compound dry and hot events. Environ. Res. Lett., 14:114034.
mp=matrix(rnorm(20,0,1),ncol=1) # 20-year precipitation of a specific month (e.g., August) mt=matrix(rnorm(20,0,1),ncol=1) # 20-year temperature of a specific month (e.g., August) threp=20 thret=80 DH<-GetDH(mp,mt,threp,thret)mp=matrix(rnorm(20,0,1),ncol=1) # 20-year precipitation of a specific month (e.g., August) mt=matrix(rnorm(20,0,1),ncol=1) # 20-year temperature of a specific month (e.g., August) threp=20 thret=80 DH<-GetDH(mp,mt,threp,thret)
Extract compound wet-hot (WH) occurrences based on thresholds of precipitation and temperature.The binary variable of the WH (or wet-warm,WW) event can be obtained.
GetWH(mp,mt,threp,thret)GetWH(mp,mt,threp,thret)
mp |
Precipitation |
mt |
Temperature |
threp |
Threshold of precipitation (e.g., 80th percentile) |
thret |
Threshold of temperature (e.g., 80th percentile) |
The occurrence of compound wet-hot events (0-1 binary variable)
Hao, Z. et al (2013). Changes in concurrent monthly precipitation and temperature extremes. Environ. Res. Lett., 8(3): 034014.
mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) threp=80 thret=80 WH<-GetWH(mp,mt,threp,thret)mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) threp=80 thret=80 WH<-GetWH(mp,mt,threp,thret)
Use the meta-Gaussian model to construct conditional distributions of the impact variable (Y) given drought and hot conditions P(Y|PRC,TEM).
ImpactMG(PRC,TEM,Y,u0)ImpactMG(PRC,TEM,Y,u0)
PRC |
Precipitation or drought indicator corresponding to the impact variable Y |
TEM |
Temperature or heat indicator corresponding to the impact variable Y |
Y |
Impact variable (e.g., Crop yield) |
u0 |
Initial condition of (PRC,TEM) |
A vector of conditional mean and variance evaluated at u0
Feng, S. et al. (2019). Probabilistic evaluation of the impact of compound dry-hot events on global maize yields. Sci. Total. Environ., 689: 1228-1234.
Hao, Z. et al. (2018). A multivariate approach for statistical assessments of compound extremes. J. Hydrol., 565: 87-94.
PRC=matrix(rnorm(60,0,1),ncol=1) TEM=matrix(rnorm(60,0,1),ncol=1) Y=matrix(rnorm(60,0,1),ncol=1) u0=c(-1.2,1.2) # Speficify the compound dry-hot condition ImpactMG(PRC,TEM,Y,u0)PRC=matrix(rnorm(60,0,1),ncol=1) TEM=matrix(rnorm(60,0,1),ncol=1) Y=matrix(rnorm(60,0,1),ncol=1) u0=c(-1.2,1.2) # Speficify the compound dry-hot condition ImpactMG(PRC,TEM,Y,u0)
Compute joint probabilities of compound dry-hot events and the independent case.
LMFDH(mp,mt,threp,thret)LMFDH(mp,mt,threp,thret)
mp |
Precipitation |
mt |
Temperature |
threp |
Threshold of precipitation (e.g., 50th percentile) |
thret |
Threshold of temperature |
Joint probability of DH divided by that of independent case
Zscheischler, J. and S. I. Seneviratne (2017). Dependence of drivers affects risks associated with compound events. Science Advances, 3(6): e1700263.
mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) threp=20 thret=80 res<-LMFDH(mp,mt,threp,thret)mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) threp=20 thret=80 res<-LMFDH(mp,mt,threp,thret)
Fit the logistic regression model (LGR) based on occurrences of compound events (Y) and climate index (CI).The output is the predicted probability of compound event occurrence for the given climate index value CI0
PredLGR(Y,CI,CI0)PredLGR(Y,CI,CI0)
Y |
Occurrences of compound dry-hot events (0-1 binary variable) (L lead time) |
CI |
Climate index (CI) as the driving factor of compound events (e.g., ENSO) |
CI0 |
Specified CI value based on which the prediction is issued |
Probability of occurrences estimated at CI0
Hao, Z. et al. (2019). Statistical prediction of the severity of compound dry-hot events based on ENSO . J. Hydrol., 572: 243-250.
CI=c(-0.7,-1.2,1.3,0.7,-0.6,1.1,-0.5,0.8,0.5,-0.5,1.6,-1.8,-0.5,-1.4,-0.1,2.2,-0.7,-1.1, 0.6, -1.7) Y=c(0,0,1,1,0,0,0,0,0,0,1,0,1,0,0,1,0,0,0,0 ) PredLGR(Y,CI,2)CI=c(-0.7,-1.2,1.3,0.7,-0.6,1.1,-0.5,0.8,0.5,-0.5,1.6,-1.8,-0.5,-1.4,-0.1,2.2,-0.7,-1.1, 0.6, -1.7) Y=c(0,0,1,1,0,0,0,0,0,0,1,0,1,0,0,1,0,0,0,0 ) PredLGR(Y,CI,2)
Compute SCEI based on monthly precipitation and temperature.
SCEI(mp,mt,ts)SCEI(mp,mt,ts)
mp |
monthly precipitation |
mt |
monthly temperature |
ts |
time scale |
The monthly SCEI series
Hao, Z. et al., 2019a. Statistical prediction of the severity of compound dry-hot events based on El Ni??o-Southern Oscillation. J. Hydrol., 572, 243-250.
mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) ts=3; # ts<=12 otherwise you should revise line 98 nd<-SCEI(mp,mt,ts) d=cbind(mp,mt,nd) testd<-matrix(d, ncol=3,byrow=FALSE)mp=matrix(rnorm(120,0,1),ncol=1) mt=matrix(rnorm(120,0,1),ncol=1) ts=3; # ts<=12 otherwise you should revise line 98 nd<-SCEI(mp,mt,ts) d=cbind(mp,mt,nd) testd<-matrix(d, ncol=3,byrow=FALSE)