Climate Change above the Cloud: Accelerating Climate. Change Research with Cloud Computing Infrastructure. Yaser Jararweh, Mahmoud Al-Ayyoub, Izzat ...
Climate Change above the Cloud: Accelerating Climate Change Research with Cloud Computing Infrastructure Yaser Jararweh, Mahmoud Al-Ayyoub, Izzat Alsmadi*, and Darrel Jenerette+ Jordan University of Science and Technology, Irbid, Jordan * Boise State University, Idaho, USA + University of California, Riverside, USA The analysis of large-scale data for the purpose of extracting patterns is applicable to several research fields. However, the size of this data is rapidly growing on a daily basis creating a need for new computing paradigms capable of handling such growing data efficiently. Cloud computing is one of the possible solutions to satisfy this pressing need. In this paper, a cloud computing based study for large scale climate related historical data from Jordan is conducted. The main focus is to accelerate the experimental part of the climate research using the cloud computing Infrastructure which will lead to faster results generation. Climate change research is receiving a lot of interest as the climate change phenomena is expected to have a direct as well as an indirect impact on human life. However, the amount of computational resources required to conduct such research in a useful and practical manner is very high. This work aims at accelerating the climate changes related research by exploiting the seemingly limitless cloud computing resources. This will help in faster results generation and accurately building climate models in reasonable time. Also, we are planning to have a cloud based portal for climate researchers who are not computing experts [1]. For the experiments and analysis, we use time series functions from different applications including R-Package-forecast and WEKA time series. Some of the time series methods are: ARIMA, ar, HoltWinters and StructTS. The initial analysis shows that ARIMA can be the best choice for our data and forecasting. Autoregressive Integrated Moving Average (ARIMA) time series algorithms or models include an explicit statistical model. It can handle irregular components of time series. This allows for non-zero autocorrelations in the irregular component. ARIMA models are defined for stationary time series. Some preprocessing methods are used to make a time series stationary. An example of those is the function (diff) in R-Package [2]. Our experimental framework is an IBM CloudBurst system. We used two virtual machines with four virtual CPUs and 32 GB of memory each. The operating system is Windows 7. Our results show that using the cloud system to conduct a computing intensive climate change experiments is very promising as we are able to generate large number of results in a relatively short time compared with using conventional machines. The results we are presenting are for climate related data from all weather stations in Jordan over a period of about 20 years. The data is collected for the years from 1990 to 2012. We present two types of results, the time series analysis results and the forecasting results that are used to predict future possible evolution in weather attributes. Based on these results, we found that humidity and dew point are the two weather attributes that showed significant increase. Such increase is categorized under global warming and is expected to cause impacts on the life of: plants, animal and human directly or indirectly.
References [1]
[2]
Jararweh, Y., Hary, A., Al-Nashif, Y. B., Hariri, S., Akoglu, A., and Jenerette D. "Accelerated discovery through integration of kepler with data turbine for ecosystem research." in: IEEE/ACS International Conference on Computer Systems and Applications, 2009. AICCSA 2009, IEEE, 2009, pp. 1005–1012. Jararweh, Y., Alsmadi, I., Al-Ayyoub, M., and Jenerette, D. "The Analysis of Large-Scale Climate Data: Jordan Case Study." in: IEEE/ACS International Conference on Computer Systems and Applications 2014.
