Agricultural Research Management Information System

Home
Research Summary
Search
About Us
- ARMIS
- Brochure
Contact Us
Report
User Request
Data Input
Help
- Operation Manual
  - PDF
  - Video
- Program Area & Commodity
We have reached 37600 number of research entries at this moment.

- Logout

Research Detail

Home
Research
Detail

Appropriateness of Probability Distributions for Analyzing Rainfall Droughts

L. R.Khan
Department of Irrigation & Water Management, Bangladesh Agricultural University,Mymensingh

Abstract/ Executive Summary:

This study deals with the probability analysis of rainfall data to ascertain the magnitude of droughts for various durations. Four different distribution namely. Normal, 2-parameter Log-Normal EVL-I and Weibull were used in this investigation. The design droughts for various durations were estimated from the probability plots. The applicability of the distributions in this drought analysis was tasted based on standard goodness of fit criteria. From the group of test using correlation. Chl-square and Kolmogoroy-Smirnoy criteria. It was confirmed that among the selected distributions for rainfall drought analysis, the best fitted method is the 2- parameter Log-normal distribution.

Keywords: Rainfall, Drought. Partial duration series, Reduced variate

Location: Mymensingh

Start Date:

End Date:

Program Area: Crop-Soil-Water Management

Commodity: Water management

Objectives:

To find the appropriateness of the selected distributions for drought analysis on the basis of some standard statistical tests,

Methodology:

Rainfall data for station number R-73 which is located at Mymensingh (latitude 24026' and Longitude 90029') was used for this study. Rainfall data for this station have been recorded by Bangladesh Water Development Board. Monthly data during the period of 1948 to 1985, except 1971 (i.e. 36 years) was collected from this organization. The selected theoretical probability distributions can be fitted to sample data either graphically or analytically. The analytical fitting of the distribution functions contain parameters which are to be estimated from sample data using some statistics. e.g. moments, maximum likelihood, least square, etc. Goodness of Fit Tests: In order to find which distribution is suitable for drought anaylsis of rainfall the following criteria of goodness of fit tests were considered: I) Correlation coefficient, ii) Chi-square Xl, and iil) Kol- mogorov-Smirnov ( ON ). The Kolmogorov-Smirnov statistics is based on the deviations of the empirical distribution function PN (x) from the completely specified continous hypothetical distribution function P (x) which is expressed as (Yevjevlcb,1972).

Source of Information: Bangladesh J. Agril. Sci. 17 (2) : 93-200: July,1990 ISSN 0379-4296

Article Link (Online Journal):

Funding Source:

Total Budget (Tk.) :

Achievement/Findings:

Out of the four selected distributions the 2-parameter Log -Normal is the best for analvzing rainfall drought series. In order to ranks from these tests, Normal distributions is the second and Weibull distribution is the third choices, respectively. Althouh this investigation was based on the rainfall data of Mymensingh only, it may be used as guideline for drought analysis of rainfall in any other places of the coun- try with certain reliability. From the group of test using correlation. Chl-square and Kolmogoroy-Smirnoy criteria. It was confirmed that among the selected distributions for rainfall drought analysis, the best fitted method is the 2- parameter Log-normal distribution.

Knowledge Management: Journal