# SPIGA: Calculation of Standardized Precipitation Index, using the... In SPIGA: Compute SPI Index using the Methods Genetic Algorithm and Maximum Likelihood

## Description

Calculate the standardized precipitation index (SPI) for monitoring drought using the technique of Genetic Algorithm (SPIGA) and Maximum Likelihood (SPIML) of a series of monthly rainfall for different time scales.

## Usage

 ```1 2 3``` ```SPIGA(Pmon, scale = 3, population = 500, maxIter = 50, plotGA = FALSE, plotCDF = FALSE) SPIML(Pmon, scale =3) ```

## Arguments

 `Pmon` monthly precipitation series ordered according to time. It is a data frame with columns: year, month, station 1, station 2, etc. `scale` an integer value representing the time scale of analysis. The most common are 1, 3, 6, 9, 12, 48, etc. `population` an integer value that sets the number of population for the use of the technique of Genetic Algorithm. `maxIter` an integer value that sets the maximum number of iterations also called cycles within the concept of Genetic Algorithm. `plotGA` optional, value Boolean default false. Shows the performance versus the number of cycles in the Genetic Algorithm. `plotCDF` optional, value Boolean default false. Shows the cumulative distribution function of each station. The graphics are monthly.

## Details

The `SPIGA` and `SPIML`, are functions to calculate the SPI using Artificial Intelligence techniques - Genetic Algorithms (GA) and numerical method - Maximum Likelihood (ML) and both provide quantitative results for monitoring DROUGHT. The GA optimize the parameters alpha and beta of the probability function Gamma given by McKee.

The `population` parameter must be an integer and balanced value, large values can generate higher time run, ie, high computational effort and small values can influence the accuracy of the results. By `plotGA` option and its corresponding graph, you can see the number of cycles to obtain a proper balance of the accuracy of the results and the computational effort.

#### Input data

similar to `Pm_Pisco`.

 Year Mon st_1 st_2 st_3 st_4 1981 1 120.25 125.25 90.55 150.25 1981 2 145.25 140.25 120.70 145.50 1981 3 120.80 150.28 90.50 130.40 1981 4 90.25 80.25 70.52 120.40 1981 5 50.25 58.25 60.50 80.50 1981 6 40.25 38.45 80.25 50.40 1981 7 20.25 30.69 50.40 40.40 1981 8 1.25 8.85 10.40 25.80 1981 9 25.25 14.25 5.80 20.80 1981 10 13.25 10.23 10.50 30.45 1981 11 50.25 40.25 30.50 80.50 1981 12 80.25 90.52 80.70 90.40 1982 1 145.80 110.25 105.40 120.25 . . . . . . . . . . . . . . . . . .

## Value

Functions `SPIGA` and `SPIML` return values saved in .txt formats (Tabular) and .pdf (graphics). They are located in the working folder of `R` [getwd()].

## Note

Dependencies: the SPIGA function, depend on the library `GA`.

## Author(s)

Iv<c3><a1>n Arturo Ayala Bizarro <ivan.ayala@unh.edu.pe>

Jessica Z<c3><ba><c3><b1>iga Mendoza <zumeje@gmail.com>

## References

McKee, Thomas B. and Doesken, Nolan J. and Kleist, John. 1993. The relationship of Drought Frequency and Duration to Time Scales. Eighth Conference on Applied Climatology

A. Belauneh and J. Adamowski. Standard Precipitation Index Drought Forecasting Using Neural Networks, Wavelet Neural Networks, and Support Vector Regression. Applied Computational Intelligence and Soft Computing, http://dx.doi.org/10.1155/2012/794061

`SPIFromParameters` to calculate the standardized precipitation index, from alpha and beta parameter of the Gamma function.

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23``` ```#### Load data data(Pm_Pisco) Pmon<-Pm_Pisco # dataframe Precipitation summary(Pm_Pisco) # view summary Pmon<-Pm_Pisco[,] #### Computing SPI with Genetic Algorithms pob <-50 # Define population number iMax <-10 # Define Max iteration # Total stations calculation. It may take some time. #SPIGA(Pmon, scale=3, population=pob, maxIter = iMax, plotGA=TRUE, plotCDF=TRUE) # station 1 computing Pmon1<-data.frame(Pmon[,1:2], Pmon\$Pm_St1) SPIGA(Pmon1, scale=3, population=pob, maxIter = iMax) # station 2 computing Pmon2<-data.frame(Pmon[,1:2], Pmon\$Pm_St2) SPIGA(Pmon2, scale=3, population=pob, maxIter = iMax) #### Computing SPI with Maximun Likelihood SPIML(Pmon, scale=3) ```

### Example output

```Loading required package: GA
Package 'GA' version 3.1.1
Type 'citation("GA")' for citing this R package in publications.
year          month            Pm_St1             Pm_St2
Min.   :1981   Min.   : 1.000   Min.   :  0.5866   Min.   :  0.5362
1st Qu.:1989   1st Qu.: 3.250   1st Qu.: 16.5649   1st Qu.: 14.0321
Median :1998   Median : 6.000   Median : 46.8745   Median : 47.4559
Mean   :1998   Mean   : 6.476   Mean   : 62.0483   Mean   : 59.3716
3rd Qu.:2007   3rd Qu.: 9.000   3rd Qu.: 98.8216   3rd Qu.: 96.8824
Max.   :2015   Max.   :12.000   Max.   :236.0222   Max.   :239.9725
Pm_St3             Pm_St4
Min.   :  0.4474   Min.   :  0.5104
1st Qu.: 13.4251   1st Qu.: 18.5945
Median : 43.8687   Median : 48.0113
Mean   : 57.3330   Mean   : 65.4650
3rd Qu.: 90.9596   3rd Qu.:100.2296
Max.   :236.6156   Max.   :244.5666
********** Calculation Station Pmon.Pm_St1

SPIGA: Successful Analysis
See the results:  /work/tmp
There were 24 warnings (use warnings() to see them)
********** Calculation Station Pmon.Pm_St2

SPIGA: Successful Analysis
See the results:  /work/tmp
There were 24 warnings (use warnings() to see them)
********** Calculation Station Pm_St1
********** Calculation Station Pm_St2
********** Calculation Station Pm_St3
********** Calculation Station Pm_St4

SPIML: Successful Analysis
See the results:  /work/tmp
```

SPIGA documentation built on May 2, 2019, 6:32 a.m.