I have temperature and precipitation data of 30 years and I'm willing to study the trend over these years of Kathmandu city. I'm very new to this study so would be really grateful if you would suggest me the statistical test and the softwares that could be used.
Mann Kendall test is ideal for estimating the trend... XLSTAT software can be used to do the MK test
You may collect useful information fromArticle Variability and time series trend analysis of rainfall and t...
, concerning the main question of the thread.Regards
use sen slope calculator to calculate the gradient of the trend/year
then use the mankendal tau to compute the trends and test the significance of the trend,
plot the annual cycle and the inter annual variability
A lot of people in data analysis use R (see https://www.r-project.org/) which is free.
I am sure there is a Mann Kendall test in R somewhere.
For Trend analysis, Sens slope, is a robust estimator. To check the significance, you can you Kendall tau estimate.
For software, I believe, R (Free and Opensource) will probably suit your need very well. It is very popular for data science and statistics. You can find all the functionality you are looking for in zyp package. To manage R codes, you may also install RStudio as suggested by Matthias Mohr .
Currently no single method is best.
Combination of method is used to assess the climate projections.
As India and Nepal are very diverse in topography in comparasion to rest of country. So here we need to be very attentive for the climate modelling.
Mann-Kendall using XLSTAT couple it to sen' s slope and you'll obtain the magnitude of change per year under the form y = sen's slope x + intercept . Replace x by your study period and multiply it by the sen's slope you have the magnitude of change over your study period. Proceed to future projections by changing the value of x
Best of luck
If you are analyzing extremes of rainfall and temperature in years, you can fit your data using the Generalized Extreme Value (GEV) distribution which gets the maximum of your element (e.g rainfall) in a year or the Generalized Pareto (GP) distribution that focuses on a certain threshold, using R or R-instant `you can easily do your analysis.
Hi for this case since you have metereological data. You need to take in count that we can analyze the mean time conditions or extreme conditions. For the first case we can use empirical cumulative distribution functions, statistical model adjusment and joint distributions and some possibilities are:
1) Fourier analysis (spectrum analysis)
2) Periodogram
3) Autocorrelation functions
4) Autocovariance functions
5) Filters (ex. Lanczos filtering)
In the other hand if you want to analyze extremes you have 4 possible methods :
1) Annual Maximums method
2) RLO. R annual maximums
3) POT. Peak over threshold
4) Method of maximums in a block of time
I hope this helps.
Well, as to me it is better if you use XLstat for both M-k and sens slop test unless using RStudio is a good fit b/c is free to use.
You can also use Nicholson Index(Ni) for rainfall and Ombrothermic diagramme
You can also use Nicholson Index(Ni) graphe and Ombrothermic Diagram.
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Climate Science
- Climate Change