This project aims to use Python programming language to build three Hydrology models and to use Mean Square Error to find the best model.
The Seattle Rainfall dataset is from kaggle:
https://www.kaggle.com/rtatman/did-it-rain-in-seattle-19482017.
It includes daily data about max temperature, min temperature and rainfall intensity from 1948-2017.
The probability is a kind of method to determine the possibility of an event occurring. A higher probability means that the event will happen more frequency.
The probability of exceedance express the chance for a specified value being exceeded in a certain cycle. The equation for probability of exceedance, such as California, Weibull, Sevruk and Geiger are listed in the Table.
Method | Equation |
---|---|
Weibull | |
Sevruk and Geiger | |
California | |
Hazen |
A return period, also known as a recurrence interval (sometimes repeat interval) is an estimate of the likelihood of an event, such as an earthquake, flood, landslide, or a river discharge flow to occur. It build from probability of exceedance and demonstrated the recurrence frequency for certain event. What' more, the return period is usually used for risk analysis to predict the hazard event.
This project will use probability of exceedance to display the probability for each maximum precipitation in selected periods and use Weibull method to calculate the probability for each rainfall events. In addition, this project will construct the return period for Seattle.
Normal Distribution, Gumbel Distribution and Log Pearson Type III Distribution will be chosen in this project to build Hydrology model.
Normal distribution is very well to established with a good sample size and sample mean of population will follow normal distribution.
Gumbel distribution is selected to perform the flood probability analysis.
Log Pearson Type III is statistical technique for fitting frequency distribution data to predict the design flood for a river at different site.
In this project, Python programming language is using to build three models and also using to compare each model. The result of Model Comparison is shown in the code in:
https://github.com/anyarcherty/Seattle_Rainfall/blob/master/Seattle%20Weather.ipynb
In this project, Log Pearson Type III model works well to find annual maximum rainfall in Seattle. For future research, my suggestion is to obtain enough precipitation data at different rainfall stations near Seattle and try to build precipitation intensity network in Seattle. Also, I suggest building the models based on the different data analysis models.