In extreme rain events (Trenberth, Weather and climate

In 1999 Trenberth found that the ultraviolet rays produced by the greenhouse effect can increase the evaporation of water and this helps to increase the capacity of the atmosphere to hold more water vapour in hot days. This is one of the main reasons for the intensification of the hydrological cycle and it increases the frequency of extreme rain events (Trenberth, Weather and climate extremes, 1999). From a research conducted in 2007 it was found that the surface temperature has increased by 0.7 C over the last century with a significant warming in several regions and another thing that found by the was the rate of the temperature increment in the land areas is higher than that in oceans (Trenberth, et al., 2007).

2.2 Rainfall patterns

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Rainfall is one of the most unpredictable climatic variables. Though it varies in an unpredictable way if we analyse rainfall data over a period of time it shows some patterns. These patterns can be varied according to the area we obtained the data and the methods we use (Douka, 2017). So the rainfall patterns can be identified spatially. For an example, in Sri Lanka, we can identify three main parts in the island according to the precipitation. Those are the wet, intermediate and dry zone (Malmgren, Hulugalla, Hayashi, & Mikam, 2003).

Figure 1: Climatic zones in Sri Lanka

Not only spatial patterns but also the temporal patterns of rainfall can also be identified. These temporal variations in rainfall occur according to how the winds are behaving in the region, location and the geographical properties of the country. For Sri Lanka, the rainfall is governed by the monsoon system and according to that system, Sri Lanka gets rainfall under four monsoons in four time periods every year. So because of these patterns, it has been easier to predict rainfall events.

But with the climate change these patterns have started to show slight deviations from their usual paths and because of that, the design rain events is being deviated from the actual rainfall events. This makes the world consider the time trends in rainfall patterns to predict or design rain events more accurately. For analysing rainfall data to identify time trends specialized methods should be used such as Mann-Kendall test, linear regression line method ETC. For the analysis, continuous series of data should be collected from the gauging stations at least up to 30 years and data should be sorted according to our objective. As an example, if we want to identify trends in annual maximums we have to sort the data and take the annual maximum rainfall values. Then the data should be represented and it will be a help for the task

2.3 Data representation

Generally, precipitation data representation is done using tables and graphs. Data can be present according to various gauging station to check trends individually.

Figure 2: Column charts

Figure 3: Tables

 

Other than this various types of graphs can be used to represent variation of rainfall properties such as intensities, depth ETC and trends with the time.

Figure 4:  Other tables

2.4 Analysing methods

Various parametric and nonparametric methods have been used to identify the trends in time series of rainfall in research work. Standard linear regression analysis and Mann Kendall test are the methods used mostly to analyse the hydro meteorological time series among them. Apart from that various other methods also have been used in order to find various relationships. As examples Shapiro-Wilk test for identifying temporal classes and climatic risks, Spearman’s rank correlation analyses to find the relationship between the events ETC.

2.4.1 Mann-Kendall test

This test is widely used in different fields of research for the detection of trends in time series because of its simplicity and its ability to deal with missing values and values below a detection limit. Mann-Kendall test equations are as follows. (Ampitiyawatta & Guo, 2010)

Where:

 

                                                                        (4)

 

Where Zmk  Mann-Kendall statistics, n is the length of the data set, Xj and Xi are sequential data values, m is the number of tied groups (a tied group is a set of sample data with the same value), and t is the number of data points in the mth  group. A positive value of Zmk indicates an increasing trend and negative value of Zmk indicate a decreasing trend.

2.4.1 Standard linear regression analysis

Following parameters are used to calculate the trend of the rainfall pattern.

Where:

  Is the slope of the trend line

Positive value means that increasing trend & negative value means that decreasing trend.

2.5 Results

Significant results on temporal trends in rainfall patterns have been obtained up to now by various research all around the world in their respective study areas. (Ampitiyawatta & Guo, 2010)Have done a research in Kalu Ganga basin to identify time trends in that basin using eight gauging stations and identified following trends in those gauging stations

Table 1: Mann-Kendall statistic values in         gauging stations

 

They have also obtained the obtained temporal distribution of monthly rainfall trends from January to December. The results imply that the Kalu Ganga basin shows a decreasing rainfall except for January.

 

 

 

 

 

 

 

Table 2: Magnitude of positive/negative and average trends for different months