Statistical Modeling of Compound Events

Description

Tools for extracting occurrences, assessing potential driving factors, predicting occurrences, and quantifying impacts of compound events (with focus on dry-hot events) in hydrology and climatology.

Details

Examples of compound events in hydroclimatology with focus on compound dry-hot events. The function GetDH is used for extracting occurrences based on thresholds of dry and hot indicators. The function ChangeDH is used for changes in compound dry and hot events. The function DriverLogit is used for assessing potential driving factors of compound events based on logistic regression model. The function PredLogit is used for predicting occurrences of compound events. The function ImpactMGDH is used for quantifying impacts of compound dry and hot events on a sector (e.g., crop yield, GPP,NDVI) based on meta-Gaussian model. The function SCEIDH is used for computing the standardized compound event index of compound dry-hot events.

Author(s)

Zengchao Hao

References

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.

Hao, Z et al. (2022). Compound droughts and hot extremes: Characteristics, drivers, changes, and impacts. Earth Sci. Rev., 235, 104241.

Obtain the accumulation of monthly hydro-climatic variables

Description

Obtain the accumulation of monthly hydro-climatic variables

Usage

ACCU(X, ts = 6)

Arguments

Examples

X=runif(120, min = 0, max = 100) # 10-year monthly data
Y<-ACCU(X,ts=3) # Compute the 3 month  accumulated series

Changes in compound dry-hot events

Description

Assess absolute/relative changes in compound dry-hot (DH) events at the monthly scale and event occurrences at the annual scale.

Usage

ChangeDH(pre,tem,threp,thret)

Arguments

Value

The absolute/relative change for each month and annual occurrence

References

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.

Examples

pre=runif(1200, min = 0, max = 100) # 100-year monthly precipitation
tem=runif(1200, min = 0, max = 100) # 100-year monthly temperature
threp=20 #Specify the percentile-based threshold
thret=80 # Specify the percentile-based threshold
chg<-ChangeDH(pre,tem,threp,thret) #  chg$abschg,chg$relchg, chg$yrtotal

Changes in compound humid-hot events

Description

Assess changes in compound humid-hot (WH) occurrences at monthly and annual scales.

Usage

ChangeWH(pre,tem,threp,thret)

Arguments

Value

The absolute/relative change for each month and annual occurrence

References

Hao, Z., AghaKouchak, A., Phillips, T.J., 2013. Changes in concurrent monthly precipitation and temperature extremes. Environ. Res. Lett. 8(3), 034014.

Examples

pre=runif(1200, min = 0, max = 100) # 100-year monthly precipitation
tem=runif(1200, min = 0, max = 100) # 100-year monthly temperature
threp=80 #Specify the percentile-based threshold
thret=80 # Specify the percentile-based threshold
chg<-ChangeWH(pre,tem,threp,thret) #  chg$abschg,chg$relchg, chg$yrtotal

Assess potential driving factors of compound dry-hot events.

Description

Use logistic regression model to establish relationship between climate indices (e.g., ENSO) and occurrences of compound dry-hot events.

Usage

DriverLogit(Y,X)

Arguments

Value

slope parameter and associated p-value

References

Hao, Z. et al. (2019). A monitoring and prediction system for compound dry and hot events. Environ. Res. Lett., 14:114034.

Examples

X=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) # climate indices
Y=c(0,0,1,1,0,0,0,0,0,0,1,0,1,0,0,1,0,0,0,0 )   # dry-hot occurrences
div<-DriverLogit(Y,X)  # div$slope and div$pvalue

Univariate empirical probability

Description

Compute univariate empirical probability

Usage

Empdis1(mp)

Arguments

Value

The empirical probability

References

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.

Examples

mp=matrix(rnorm(120,0,1),ncol=1)
ep1<-Empdis1(mp)

Bivariate empirical probability

Description

Compute bivariate empirical probability

Usage

Empdis2(mp,mt)

Arguments

Value

The bivariate empirical probability

References

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.

Examples

mp=matrix(rnorm(120,0,1),ncol=1)
mt=matrix(rnorm(120,0,1),ncol=1)
ep2<-Empdis2(mp,mt)

Occurrence of compound dry-hot events

Description

Get compound dry-hot (DH) events based on thresholds precipitation and temperature, resulting in a 0-1The binary variable

Usage

GetDH(mp,mt,threp,thret)

Arguments

Value

The occurrence of compound dry-hot events (0-1 binary variable)

References

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.

Examples

mp=matrix(rnorm(60,0,1),ncol=1) # prec. of a month (e.g., August,60 years)
mt=matrix(rnorm(60,0,1),ncol=1) # temp. of a  month (e.g., August,60 years)
threp=20 #Specify the percentile-based threshold
thret=80 #Specify the percentile-based threshold
DH<-GetDH(mp,mt,threp,thret) # 0-1 series

Occurrence of compound wet-hot events

Description

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.

Usage

GetWH(mp,mt,threp,thret)

Arguments

Value

The occurrence of compound wet-hot events (0-1 binary variable)

References

Hao, Z. et al (2013). Changes in concurrent monthly precipitation and temperature extremes. Environ. Res. Lett., 8(3): 034014.

Examples

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)

Impacts of concurrent droughts and hot extremes

Description

Use meta-Gaussian model to construct conditional distributions of the standardized impact variable (SII) given SPI and STI (represented by random variables of Z, X, and Y). The output include (1) the conditional mean and variance of STI given SPI and STI (e.g., X0=-1.2 and y0=1.2) (2)The conditional probability of P(SII<=z0|SPI<=x0, STI>y0), or lower GPP given dry-hot conditions

Usage

ImpactMGDH(mp,mt,mi,x0,y0,z0=NULL)

Arguments

Value

(1) Conditional mean/variance at SPI=x0, STI=y0, here z0=NULL (2)conditional probability of lower SII given lower SPI and higher STI or P(SZI<=z0|SPI<=x0, STI>y0), here z0 should be specified

References

Feng S.,Hao Z., 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.

Examples

mp=matrix(rnorm(60,0,1),ncol=1) # precipitation of a month (60 years)
mt=matrix(rnorm(60,0,1),ncol=1) # temperature of a month  (60 years)
mi=matrix(rnorm(60,0,1),ncol=1)   # impact of a month (60 years)
imp1<-ImpactMGDH(mp,mt,mi,-1.2,1.2,NULL) # imp1$mu, imp1$sig
imp2<-ImpactMGDH(mp,mt,mi,-1.2,1.2,0)# imp2$Condprob

Likelihood multiplication factor (LMF) or probability multiplication factor (PMF) of compound dry-hot events

Description

Compute joint probabilities of compound dry-hot events and the independent case.

Usage

LMFDH(mp,mt,threp,thret)

Arguments

Value

Joint probability of DH divided by that of independent case

References

Zscheischler, J. and S. I. Seneviratne (2017). Dependence of drivers affects risks associated with compound events. Science Advances, 3(6): e1700263.

Examples

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)

Prediction of compound event occurrences

Description

Fit the logistic regression model (LGR) based on occurrences of compound events (Y) and climate index (CI).The output is predicted probability of compound event occurrence for the given climate index value CI0 or the prediction performance (using AUC) and predicted values based on LOOCV

Usage

PredLogit(Y,X,type,x0=NULL)

Arguments

Value

(1) Probability of occurrences estimated at x0 if type="prediction", shouLd input x0 (2) AUC values and predictions LOOCV if type="performance", x0 is not needed

References

Hao, Z. et al. (2019). Statistical prediction of the severity of compound dry-hot events based on ENSO . J. Hydrol., 572: 243-250.

Examples

Y=as.matrix(c(0,0,1,1,0,0,0,0,0,0,1,0,1,0,0,1,0))
X=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) #  dry-hot (or other) 0-1 events
pred1<-PredLogit(Y,X,type="prediction",2) #pred1$predx0
pred2<-PredLogit(Y,X,type="performance") #pred2$aucvalue,pred2$predictions

Standardized Compound Event Indicator (SCEI) for dry-hot events

Description

Compute SCEI based on monthly precipitation and temperature.

Usage

SCEIDH(pre,tem,ts)

Arguments

Value

The monthly SCEI series of several decades

References

Hao, Z. et al., 2019. Statistical prediction of the severity of compound dry-hot events based on El Ni??o-Southern Oscillation. J. Hydrol., 572, 243-250.

Examples

pre=matrix(rnorm(120,0,1),ncol=1) # 10-year monthly precipitation
tem=matrix(rnorm(120,0,1),ncol=1) # 10-year monthly temperature
ts=1; # ts<=12 otherwise you should revise the function
SCEIDH(pre,tem,ts) #Generate monthly data

Monthly precipitation (pre), temperature (tem), Niño3.4 (nino) and GPP(gpp)

Description

A dataset of monthly precipitation, temperature, and Niño3.4 index for 1951-2016 and monthly GPP for 1981–2016 (1951-1980 are NAN).

Usage

data(dfptng)

Format

A data frame with 792 rows (12 months × 66 years) and 4 variables:

pre

Monthly precipitation(unit:mm/month)

tem

Monthly average temperature (unit: °C)

nino

Monthly ENSO index (e.g., Nino3.4)

gpp

GPP

Source

The data were obtained from ERA-5, NOAA, and FLUXCOM.