Decision Support System for Cyclone and Weather Forecasting

NB: We do not resell your papers. Upon ordering, we do an original paper exclusively for you.

NB: All your data is kept safe from the public.

Table of Contents

Executive Summary
Introduction
Current Practices
Analysis and Design
Feasibility
SWOT
Conclusion
Reference List

Executive Summary

Decision Support System tools are critical for making robust and reliable decisions. Predictive models could enhance decisions and improve future outcomes. Therefore, this proposal presents a predictive time-series model as a Decision Support System for cyclone forecasting in Northern Queensland and other areas in Australia, which are prone to cyclones. The model would rely on readily available historical cyclone records and current data to forecast future outcomes on a long-term basis. It would also account for future weather patterns. Insights derived from the data would help farmers to make informed decisions regarding their farm operations. In the past, farmers have incurred massive losses because they relied on traditional, historical patterns, and normal methods of cyclone and weather forecasts. However, these patterns would not be necessarily similar in the future. Therefore, a predictive model can provide insights that farmers require for better decision-making. Adoption of such a model would enhance efficiency, help farmers to avoid expensive investments that cyclones would damage, save costs, and improve the public good.

Introduction

A Decision Support System (DSS) is an application platform for decision-making (Decision Support System, 2009). It relies on data modelling, communication technologies, knowledge, and documents to recognise and solve issues, run decision processes, and make decisions (Power, 2014). DSS computer applications enhance decision-making capabilities of an individual or a group. DSSs may also reflect academic areas of research for designing and studying analytical information processes.

Therefore, the proposal focuses on how such a system can enhance decision-making capabilities for farmers if the Bureau of Meteorology, Australia adopts it. The DSS shall be a data-driven system for analytical and data modelling purposes.

In the past, Cyclone Larry had hit hard Northern Queensland with devastating results to farmers and residents. Northern Queensland has a wide “sugar belt and other tropical horticultural crops like bananas” (United States Department of Agriculture, 2006). The region has over 8,500 farmers. Cyclones result into massive damages, including destruction of crops and farm equipment (United States Department of Agriculture, 2006). Northern Queensland accounts for Australia’s banana industry and over 25% of the country’s sugar cane. Some reports indicated that Cyclone Larry had destroyed over 200,000 tonnes of bananas, worth $ASD 300 million or $USD 215 million (United States Department of Agriculture, 2006).

Consequently, farmers live in constant fears of cyclones. Although cyclone is a complex natural process with forces that exceed human control, science and technologies can help in predicting its patterns and allow farmers to make informed decision about planting and investing in farm machinery.

Current Practices

The Bureau of Meteorology provides current patterns of cyclones in Australia. However, the organisation does not provide long-term forecasts for cyclone and weather patterns. The Bureau can only predict cyclones at 12, 24, and 48 hour time-steps (Bureau of Meteorology, 2014). Although short-term weather forecasts are suitable for farmers on their day-to-day farming activities (IBM Research, n.d), there is a need to address several issues regarding changes in weather patterns due to global warming. For instance, a study by Haig, Nott, and Reichart (2014) noted that “the number of tropical cyclones hitting Queensland and Western Australia has fallen to low-levels not seen for more than 500 years” (p. 667).

However, the current statistical methods of weather prediction have drawbacks. These methods involve collection of statistical data about different weather patterns at a given period of the season. Afterwards, analysts show results on charts. The charts indicate averages of highest and lowest possible weather patterns based on prevailing circumstances of the day. Once analysts depict information on charts, they presume that weather patterns will follow previous sequences in the future. Such simple assumptions can no longer work with the dynamic and unpredictable weather patterns. Moreover, they cannot account for a large number of variables based on big historical data on cyclone patterns. Past and future weather patterns are no longer identical, as traditional forecasting methods had depicted.

Analysis and Design

The proposed approach shall rely on predictive big data-driven algorithms to provide cyclone forecasts, including other weather patterns for farmers. The proposed analytic approach will use data from tropical cyclone records, which are less than 50 years based on the available records. The analysis shall review past records and current patterns to derive long-term insights for future cyclone patterns with lead times of several years.

This is big data. It would involve several calculations that entail hundreds of cyclone patterns collected over the last 50 years. The model will provide quantitative probabilistic forecasts that would be able to depict correlation between past cyclone patterns, present observational patterns, and long-term forecasts.

The specific predictive model would be time-series forecasting. This would involve specific variables, which have depicted past changes. Forecasts from data-driven approaches are credible because the core data have distinct historical trends based seasonality (Bala, 2012).

Cyclone variables would include changes in intensity or course, speed, and specific environmental patterns, such as temperature, cloudiness, wind rate, direction, season, and humidity at the time of occurrence. The predictive approach would account for temperatures and other weather variables.

Tools for the project would include robust analytic software, such as R, SPSS, and SAS (Stanton and De Graaf, 2013). However, the choice of specific software shall depend on its availability.

Farmers would adopt insights from data analysed to minimise losses incurred because of cyclones and other extreme weather patterns, such as drought and flood.

Feasibility

The project is highly feasible, and other sectors have already adopted big data predictive modelling to guide their future operations.

SWOT

Strength

The predictive model would offer great values on long-term cyclone and weather forecasts
The techniques are highly robust with deep analytical skills
Some analytical tools are open source
Farmers will have comprehensive understanding of cyclones and weather impacts on their farm operations

Weaknesses

Some analytic tools are extremely expensive
Failure to account for critical variables may lead to poor outcomes
Predictive model may not always be accurate (there is a confidence level)

Opportunities

Predictive models are applicable in other fields
It is a growing area with a promising future
Abundant data from past records

Threats

There are no adequate data scientists for the job
Expensive software renders the model impractical for many potential users
Data may be inconsistent

Conclusion

The ability to forecast cyclones, weather patterns, and associated risks on a long-term basis is an idea that farmers have long sought. The project outcomes would protect farmers from risks, enhance profitability, and public good. By relying on robust, new predictive analytic models to forecast cyclones and weather risks, farmers will gain significant edge in their operations and markets. Moreover, the model would enhance efficiency and help farmers to avoid planting and investing in farm inputs during downtime caused by unpredictable adverse cyclone and weather patterns. This would reduce costs and save resources.

Reference List

Bala, D 2012, Time Series Forecasting: choosing from data driven vs. model based methods, Web.

Bureau of Meteorology 2014, Tropical Cyclone Forecasting, Web.

Decision Support System 2009, Web.

Haig, J, Nott, J, and Reichart, G-J 2014, ‘Australian tropical cyclone activity lower than at any time over the past 550–1,500 years’, Nature, vol. 505, pp. 667–671. Web.

IBM Research n.d., Weather modeling and data analytics empower an island nation to save its natural resources, Web.

Power, D J 2014, DSS Basics, Web.

Stanton, J and De Graaf, R 2013, An Introduction to Data Science, Syracuse University, Syracuse, NY.

United States Department of Agriculture 2006, Cyclone Larry Lashes Northeastern Queensland, Web.

Click Here To Order Now!