The objective of preliminary statistical analysis in hydrology is to examine the independence and stationarity of the data series. Rao and Hamid (2000) identified four major analysis which is independence and stationarity test, homogeneity test, and the outliers and trends test that required in hydrology analysis.
The test for independence and stationarity can be conducted using Wald-Wolfowitz method (WW-Test. The equations of this test is shown below:
The homogeneity test was used to test the hypothesis of two samples data. The test is conducted using The Mann and Whitney method.
The outliers test is important to smooth the data fitting distribution. Low and high outliers will give different effects on the analysis. Grubbs and Beck test is used to detect outliers based on the assumption of normality.
The distribution test is conducted to understand the behavior of hydrology time series data.The Mann-Kendall test is widely used to detect trends in hydrologic data. The Wilcoxon inversion Test (Kanji, 1993) is used to test if two random samples have the same frequency distribution.
Showing posts with label Statistical Methods in Hydrology. Show all posts
Showing posts with label Statistical Methods in Hydrology. Show all posts
Data for Hydrology Analysis
The magnitude of data (i.e. daily, 1-day, monthly, annual) is important in hydrology and water resources analysis. In hydrology analysis those data were classified based on time base and event base data. The length of data also play important role in hydrology analysis. Based on WMO, 30 years of data is sufficient to quantify the changes in hydrology variables.
In climate change analysis, 30 years of high reliability daily precipitation is needed and for Malaysia condition data range from 1961 to 1990 is selected. Previous study shows that this range of data is good enough to investigate the current and future climate change.
Observed large-scale NCEP (National Centre for Environmental Prediction) reanalysis atmospheric variables for the same time period (1961 to 1990) have been used as predictors. The NCEP/NCAR Reanalysis produced a retroactive 51-year (1948–1998)record of global atmospheric fields derived from a Numerical Weather Prediction model kept unchanged over the analysis period and constrained by observations (Kalnay et al., 1996). NCEP daily global analyses data provided by the NCEP/NCAR internet site http://dss.ucar.edu/pub/reanalysis/.
The GCM simulations used for this study are from Hadley Centre 3rd generation (HadCM3) coupled oceanic-atmospheric general circulation model (Wilby et al., 2001). The Hadley circulation provides a useful framework for understanding the nature of large scale flow, the actual circulation in the tropics involves substantial zonal and regional variations (Manton and Bonell, 1995). This data is available at http:// www. cics. uvic.ca /scenarios / sdsm/ select. cgi. These experiments forced with changes in greenhouse gas (GHG) concentrations alone and those forced with greenhouse gas and sulphate aerosol changes. In terms of precipitation, control runs from HadCM3 transient simulations indicate approximately a 3% increase in global precipitation by the end of the 21st century (IPCC, 2001). The GCM data from 1961 to 2099 were extracted for 30-year time slices. For consistency description the scenarios data will be named as follow; the baseline period, 1961-1990 (current observed), 2010 to 2039 (the 2020s), 2040 to 2069 (the 2050s) and 2070 to 2099 (the 2080s). The justification for this division was based om a substantial change in rainfall. This point also coincide with a change justified by WMO. In climate change analysis, it is important that equal time segments are used for contrast and comparison purposes.
In climate change analysis, 30 years of high reliability daily precipitation is needed and for Malaysia condition data range from 1961 to 1990 is selected. Previous study shows that this range of data is good enough to investigate the current and future climate change.
Observed large-scale NCEP (National Centre for Environmental Prediction) reanalysis atmospheric variables for the same time period (1961 to 1990) have been used as predictors. The NCEP/NCAR Reanalysis produced a retroactive 51-year (1948–1998)record of global atmospheric fields derived from a Numerical Weather Prediction model kept unchanged over the analysis period and constrained by observations (Kalnay et al., 1996). NCEP daily global analyses data provided by the NCEP/NCAR internet site http://dss.ucar.edu/pub/reanalysis/.
The GCM simulations used for this study are from Hadley Centre 3rd generation (HadCM3) coupled oceanic-atmospheric general circulation model (Wilby et al., 2001). The Hadley circulation provides a useful framework for understanding the nature of large scale flow, the actual circulation in the tropics involves substantial zonal and regional variations (Manton and Bonell, 1995). This data is available at http:// www. cics. uvic.ca /scenarios / sdsm/ select. cgi. These experiments forced with changes in greenhouse gas (GHG) concentrations alone and those forced with greenhouse gas and sulphate aerosol changes. In terms of precipitation, control runs from HadCM3 transient simulations indicate approximately a 3% increase in global precipitation by the end of the 21st century (IPCC, 2001). The GCM data from 1961 to 2099 were extracted for 30-year time slices. For consistency description the scenarios data will be named as follow; the baseline period, 1961-1990 (current observed), 2010 to 2039 (the 2020s), 2040 to 2069 (the 2050s) and 2070 to 2099 (the 2080s). The justification for this division was based om a substantial change in rainfall. This point also coincide with a change justified by WMO. In climate change analysis, it is important that equal time segments are used for contrast and comparison purposes.
Subscribe to:
Posts (Atom)