Using IBM Watson Analytics Tool To Analyze Rainfall Variability And Predict Future Precipitation

Background Information

Because of the headway of technology and the data innovation, a gigantic sum and assortment of information are being produced each day. However, research has shown that such an immense volume and speed of information is futile until they are formed and used and carter as data (Tozer et al. 2017). The critical change in the socio economic aspects has gain importance and information are being gathered from different point for comprehension and assessing the patter of the environmental change. The manual examination of such an immense volume of information created each moment is impractical and frequently prompt different oversights and abnormalities in the investigation (Ashcroft & Karoly, 2016). The headway of technology thus have offered ascend to the real time analysis tools that aides in examining and visualizing data inside seconds. Dash boarding of such information helps the individual without no earlier learning of the data analysis method can have clear comprehension of the outcomes and data acquired. In this task, IBM Watson Analytics Tool has been chosen for assessing the precipitation inconstancy measurement and predict next 12 month rain fall at Adelaide Station in Australia.

The use of IBM Watson investigation devices furnishes the client with the real-time cloud based technology for examining a lot of information together. The Rapid change in the atmosphere and ecological condition has noteworthy effect on the agribusiness business (Imteaz et al. 2015). This specific venture in at investigating the effect of rainfall variability on the cotton production. At the point when a lot of rain falls, one of the essential issues that producers face is supplement filtering (Tozer et al. 2015). Certain supplements tend to drain more than others. "Nitrogen, potassium, sulfur and boron have a higher propensity to be filtered out of the dirt," (Kamruzzaman et al. 2015). An absence of any of these supplements can stunt a plant's development.

An excess of water can likewise leave the dirt waterlogged, which may build danger of compaction. Furthermore, oxygen in the dirt winds up noticeably exhausted following a couple days submerged. "Producers need to keep an eye out for waterlogging and oxygen exhaustion in the dirt in high precipitation years," (Contreras et al. 2013).

Rain can defer planting, which turns into an issue for yields planted early, similar to corn. "Since corn should be planted early, an excessive amount of rain in the spring that causes postponements can adversely influence the corn harvest's yield," (Tibby et al. 2016).

Reporting and Dashboard

With cotton specifically, rain toward the finish of the season can have hurtful impacts. Hardlock and boll spoil can turn out to be more common with overabundance rain, as cotton needs sunny climate for the bolls to open up before collect. Hence, it is necessary to identify the period when the variability of rainfall will be stable and the proportion of rainfall will be minimum.

The IBM Watson tool has been employed to evaluate the rainfall variability data in Adelaide over the last 10 years period from March 2006 to April 2016. In this report, the raw data has been collected from the Australian Government, Bureau of Meteorology's legitimate site. For the examination of the precipitation changeability, month to month precipitation information has been acquired from four weather station of Adelaide area. Notwithstanding that, Brookglen Farms Pty Ltd, a cotton producer firm has been chosen for giving significant proposal to acquiring the vital objectives through the advancement of the cotton production.

The examination have been led in view of the precipitation exercises seen over the Adelaide area in Australia. The points of interest of the dashboard are given in the underneath section.

Figure: Yearly Rainfall Trend Dashboard 

Figure: Dashboard on Predictive Model for all the stations

Recent research work related to the impacts of environmental change on cotton generation normally concentrates on the impact of some collected measure of rainfall. Gwimbi and Mundoga (2010) measured effect of environmental change for the whole developing period of cotton and found that cotton creation levels declined as rainfall variability diminished and temperature expanded. Along these lines the forecast of the precipitation example and inconstancy throughout the following 12 months will permit the Brookglen Farms Pty Ltd Organization in building up a strong vital arrangement for upgrading the proficiency of development and yield generation. For the dashboard formation of the prescient model, the precipitation information are utilized demonstrating the month to month normal precipitation assembled in millimeters. From the investigation it has been watched that the time of October to December have been anticipated to have the most reduced precipitation throughout the following year. The information was broke down in view of the prescient model for distinguishing the example. In view of the perception gotten from the prescient model, the association will have the capacity to mastermind elective answers for guaranteeing the water supply to the yields.

Figure: Rainfall Variability in Adelaide

Dashboard on Predictive Models

The precipitation at Adelaide is around 1530 mm every year, with around 70–80% of this falling in the winter months amongst April and October. The information utilized here with the help of IBM Watson analytics tool. Cornish comprised of four quater sums for every year, with December 31 of the earlier year being incorporated into non-jump years.

A nitty gritty investigation of the precipitation of Adelaide has built up in the above dashboard that intermittent changes happen in the occurrence and length of the winter downpours. These progressions have a period and adequacy of around 10 years and 3 months individually, and superimposed on them is a long haul slant which is showed by protraction of the last 50% of the season, spring downpours now happening around 3 weeks after the fact than they did a little more than 10 years prior. The aggregate amount of rain encouraged has demonstrated no measurably huge changes.

Figure: Rainfall Variability in North Adelaide

The above dashboard has shown that there has been a noteworthy change in precipitation designs since the 2006, with substantial geographic variety. North Adelaide has seen a huge increment in yearly precipitation, though the majority of the eastern seaboard and south-west Australia have seen a noteworthy diminishing (Tibby et al. 2016). Precipitation changes over the more extended period from 2006 to 2016 are by and large positive and are biggest in the north-west. The decline in autumn rainfall in North Adelaide has strong qualitative similarities with the decline observed in the same period in Adelaide.

Figure: Rainfall Variability in Kent Town

This particular dashboard is demonstrating the reliance of outrageous month to month precipitation at various terms of aggregation, on month of the year and station has been measured for a locale of Kent Town close Adelaide. There is confirmation that outrageous precipitation relies on upon the station and the time of the year and that the reliance contrasts for various terms of aggregation. There is no persuading confirmation of a direct pattern in the mean estimation of extremes over the 10 year time span in this district or of an adjustment in regular example. In differentiate, there is solid confirmation of an expansion in inconstancy, evaluated as a 58% expansion in total estimation of deviation from the mean.

Figure: Rainfall Variability in Keswick

The above dashboard has shown diminishing patterns of precipitation profundity at Keswick station, to be specific the rainfall variability for June and July precipitation beginning in the 2006s. No noteworthy patterns were found over the four quarter precipitation information. The staying station demonstrated expanding patterns of month to month precipitation profundity. It was also found to clarify the expanding patterns for the Adelaide (June) and north Adelaide (April) precipitation information and the diminishing patterns for Keswick (July) precipitation.

Dashboards considering the rainfall variability in against each quarter

Figure: Yearly Trend

Australian precipitation has expanded somewhat over the past century, and more so in summer than winter (). On a landmass wide premise, this pattern is not measurably critical as a result of high inter annual inconstancy. On a territorial and occasional premise, inclines in precipitation are clearer. Yearly aggregate precipitation has ascended by around 15% in NSW, South Australia (SA), Victoria and the Northern Territory (NT), with little change in the other states. South-west WA has turned out to be 25% drier in winter, with a large portion of the decrease in the vicinity of 2006 and 2008 (Tibby et al. 2016).

Higher Australian precipitation since 2006 is connected to increments in both substantial precipitation occasions and the number of rain days, with some provincial exemptions (Tozer et al. 2017). By and large, the number of wet days has expanded by around 10% (regardless of the huge 10% decrease in south-west WA), despite the fact that this figure ascends to about 20% in parts of NSW and NT (Tozer et al. 2015). Critical increments of substantial precipitation occasions in summer, particularly in the east and north, and reductions in south-west WA have happened (Tozer et al. 2015). In focal Australia, remaking of old surge successions from dregs stores in the crevasse of the Finke River demonstrate that four of the eight biggest surge occasions in the most recent 800 years have happened since 2006 (). Provincial reductions in precipitation in western Victoria have clearly influenced the hydrological spending plans of a few encased lakes (Tozer et al. 2017). Levels of three lakes have fallen 15–20 m since the 2006s.

High precipitation conveyance, generally over the cotton belt of northern and eastern Australia and a few parts of South Australia, is probably going to fundamentally effect Australia’s cotton production. The impacts of the high precipitation will be especially articulated Western Australia and South Australia, where the yield is completely reliant on rainstorm downpours.

Despite the fact that cotton production in the significant surplus conditions of NSW, QLD furthermore, Norther part is generally flooded, the yield is as yet reliant on storm downpours for recharging repositories and ground water saves required for water system and producing power to run tube wells.

A correlation of the current year's precipitation design with recorded information demonstrates that the circumstance this year to some degree like 2006 when precipitation lack amid June 1 to July 24 was 24 percent beneath typical (dashboard above). Despite the fact that the precipitation circumstance enhanced amid the second 50% of the rainstorm season in 2002, trim misfortune was huge, with cotton creation declining by 20.7 million tons from the earlier year's levels. In spite of the fact that the land circulation of precipitation insufficiency this year is somewhat unique in relation to 2006, a noteworthy decrease in cotton production this year, especially in the conditions of NSW and QLD, seems unavoidable. In any case, it is too soon to evaluate the potential generation misfortune.

As indicated by the dashboard mentioned above, dynamic cotton production is right now falling behind last year's level by more than six million hectares, which would convert into a generation loss of at slightest 12 million tons. Bring down cotton yields because generally and inconsistent storm rains in a few states would likewise bring about extra generation misfortunes. As the window of chance for planting of cotton will be over soon, agriculturists will begin moving to less water system concentrated brief length beats and coarse grains. In spite of the fact that the legislature would try full scale endeavors to lessen misfortunes by giving different impetuses and info endowments to ranchers, formulating possibility arrangements to increment of cotton production amid the winter season, an general loss of no less than 11 million tons in 2016 cotton generation seems likely. In a most noticeably awful case situation, the misfortunes could be as high as 15 million tons from a year ago's record creation of 99.15 million tons (overhauled). Nonetheless, a clearer picture will rise just by end-August, at the point when the legislature gets definite reports from different dry spell influenced states.

Figure: 12 Month Forecast

The above dashboard has shown the next 12 months forecast for each of the four location.

The generation and development of the cotton required less supply of water. Customarily, the cotton is developed amid the times of October to December with the supply of less precipitation. The cotton cultivators use the accessibility of precipitation and water supply for augmenting the cotton generation (Dijk et al., 2013). The present patterns in precipitation has brought about the diminished efficiency of the harvests bringing about the low gainfulness to the association. The pattern in the precipitation design in different locale fluctuates essentially in light of the atmosphere and climate condition (Anwar et al., 2015). The examination module created on IBM Watson Analytics devices have furnished with the subtle elements and understanding of the precipitation design. In light of the investigation and assessment different proposals have been made for upgrading the productivity and harvest creation in Brookglen Farms Pty Ltd organization. The different suggestions are as per the following:

Development time turn: Traditionally, in the Adelaide Region, the cotton are sowed amid the time of October. From the investigation, it has been watched that the precipitation amid the time of October have essentially diminished throughout the years due the adjustment in atmosphere. The reduction of precipitation have brought about the lessened creation of the cotton. Then again, from the examination it has additionally watched that the time of November and December have gotten reduced precipitation over the most recent ten years when thought about the previous years. Moreover from the prescient model, it have been examined that for the following year, the time of October has been anticipated to get less precipitation that is great for cotton development. Additionally, the development of cotton should be possible in any season dissimilar to the lasting products situated in the plenteous accessibility of water (Qureshi, Hanjra, and Ward, 2013). The less supply of water will help in guaranteeing the quality and high profitability of cotton. In this way, it has been prescribed to switch the cotton development amid the times of October and December.

Exchange Source of Water: The shortage of the water amid the development time can be killed while guaranteeing a substitute wellspring of water (Kirono et al., 2016). Generally, the cotton creation was exclusively subject to the water accessibility from the precipitation and the substitute Murrum bidgee and Murray River. The Brookglen Farms Pty Ltd needs to distinguish a close-by water hotspot for watering he trim amid the time of water inadequacy. It is additionally prescribed to spare the water amid the stormy season and reuse the water for reusing the water and effectiveness using the yearly precipitation The association needs to put resources into the use of the water discussion strategy amid the executed amid the other flooding season and drying season for guaranteeing the preservation of water amid the cotton development period.

Nonstop Monitoring: From the examinations of the information, it has been watched that because of the expansion of a risky atmospheric deviation and consistent change in temperature has been seen (Thornton et al.,2014). The Brookglen Farms Pty Ltd needs to consistently screen and investigations the adjustment in the precipitation changeability all through every one of the months in the year for guaranteeing the ideal generation of cotton and accomplishing the vital objectives.

Reference:

Aggarwal, M., & Madhukar, M. (2016). IBM’s Watson Analytics for Health Care: A Miracle Made True. Cloud Computing Systems and Applications in Healthcare, 117-134.

Anwar, M. R., Li Liu, D., Farquharson, R., Macadam, I., Abadi, A., Finlayson, J., ... & Ramilan, T. (2015). Climate change impacts on phenology and yields of five broadacre crops at four climatologically distinct locations in Australia. Agricultural Systems, 132, 133-144.

Ashcroft, L., & Karoly, D. J. (2016, April). Understanding extreme rainfall events in Australia through historical data. In EGU General Assembly Conference Abstracts (Vol. 18, p. 266).

Challinor, A. J., Watson, J., Lobell, D. B., Howden, S. M., Smith, D. R., & Chhetri, N. (2014). A meta-analysis of crop yield under climate change and adaptation. Nature Climate Change, 4(4), 287-291.

Chen, Y., Argentinis, J. E., & Weber, G. (2016). IBM Watson: how cognitive computing can be applied to big data challenges in life sciences research. Clinical therapeutics, 38(4), 688-701.

Christodoulakis, C., Kandogan, E., Terrizzano, I. G., & Miller, R. J. (2017, March). VIQS: Visual Interactive Exploration of Query Semantics. In Proceedings of the 2017 ACM Workshop on Exploratory Search and Interactive Data Analytics (pp. 25-32). ACM.

Contreras, S., Santoni, C. S., & Jobbágy, E. G. (2013). Abrupt watercourse formation in a semiarid sedimentary landscape of central Argentina: the roles of forest clearing, rainfall variability and seismic activity. Ecohydrology, 6(5), 794-805.

Dijk, A. I., Beck, H. E., Crosbie, R. S., Jeu, R. A., Liu, Y. Y., Podger, G. M., ... & Viney, N. R. (2013). The Millennium Drought in southeast Australia (2001–2009): Natural and human causes and implications for water resources, ecosystems, economy, and society. Water Resources Research, 49(2), 1040-1057.

Gandomi, A., & Haider, M. (2015). Beyond the hype: Big data concepts, methods, and analytics. International Journal of Information Management, 35(2), 137-144.

Guidi, G., Miniati, R., Mazzola, M., & Iadanza, E. (2016). Case Study: IBM Watson Analytics Cloud Platform as Analytics-as-a-Service System for Heart Failure Early Detection. Future Internet, 8(3), 32.

Haws, D. C., Rish, I., Teyssedre, S., He, D., Lozano, A. C., Kambadur, P., ... & Parida, L. (2015). Variable-selection emerges on top in empirical comparison of whole-genome complex-trait prediction methods. PloS one, 10(10), e0138903.

Heinemann, A. B., Barrios-Perez, C., Ramirez-Villegas, J., Arango-Londoño, D., Bonilla-Findji, O., Medeiros, J. C., & Jarvis, A. (2015). Variation and impact of drought-stress patterns across upland cotton target population of environments in Brazil. Journal of experimental botany, 66(12), 3625-3638.

Hoyt, R. E., Snider, D., Thompson, C., & Mantravadi, S. (2016). IBM Watson Analytics: Automating Visualization, Descriptive, and Predictive Statistics. JMIR Public Health and Surveillance, 2(2).

Hurley, B. (2015). TIM Lecture Series-Improving the Self-Service Customer Experience: The Case of IBM Watson and Purple Forge. Technology Innovation Management Review, 5(9), 36.

Imteaz, M. A., Paudel, U., & Gato-Trinidad, S. (2015). Climatic variability of potential rainwater savings: A case study for Adelaide. 9th International Water Sensitive Urban Design (WSUD 2015), 233.

Kamruzzaman, M., Aryal, R., Beecham, S., Mulcahy, D., Metcalfe, A. V., Slattery, S., & Lee, S. S. (2015). PM10 Dispersion in Adelaide and Its Relationship with Rainfall. Water, Air, & Soil Pollution, 226(12), 400.

Kirono, D. G., Butler, J. R., McGregor, J. L., Ripaldi, A., Katzfey, J., & Nguyen, K. (2016). Historical and future seasonal rainfall variability in Nusa Tenggara Barat Province, Indonesia: implications for the agriculture and water sectors. Climate Risk Management, 12, 45-58.

Lak, P., Sadat, M., Barrelet, C. J., Petitclerc, M., Miranskyy, A., Statchuk, C., & Bener, A. B. (2016). Preliminary Investigation on User Interaction with IBM Watson Analytics.

Miller, J. D. (2016). Learning IBM Watson Analytics. Packt Publishing Ltd.

O'Leary, D. E. (2017). Emerging White-Collar Robotics: The Case of Watson Analytics. IEEE Intelligent Systems, 32(2), 63-67.

Qureshi, M. E., Hanjra, M. A., & Ward, J. (2013). Impact of water scarcity in Australia on global food security in an era of climate change. Food Policy, 38, 136-145.

Ray, D. K., Gerber, J. S., MacDonald, G. K., & West, P. C. (2015). Climate variation explains a third of global crop yield variability. Nature communications, 6.

Self, R. (2016). Learning to program using immersive approaches: A case study learning SAS®, IBM Bluemix and Watson Analytics. Proceedings of the iLRN 2016 Santa Barbara.

Thornton, P. K., Ericksen, P. J., Herrero, M., & Challinor, A. J. (2014). Climate variability and vulnerability to climate change: a review. Global change biology, 20(11), 3313-3328.

Tibby, J., Barr, C., McInerney, F. A., Henderson, A. C., Leng, M. J., Greenway, M., ... & McNeil, V. (2016). Carbon isotope discrimination in leaves of the broad?leaved paperbark tree, Melaleuca quinquenervia, as a tool for quantifying past tropical and subtropical rainfall. Global change biology, 22(10), 3474-3486.

Tozer, C. R., Kiem, A. S., & Verdon?Kidd, D. C. (2017). Large?scale ocean?atmospheric processes and seasonal rainfall variability in South Australia: potential for improving seasonal hydroclimatic forecasts. International Journal of Climatology.

Tozer, C., Kiem, A. S., & Verdon-Kidd, D. (2015). Establishing the hierarchy of influence of drivers of seasonal rainfall variability in South Australia to inform seasonal rainfall forecasting. In 36th Hydrology and Water Resources Symposium: The art and science of water (p. 1561). Engineers Australia.

Tsoi, K. K., Chan, F. C., Hirai, H. W., Leung, G. K., Kuo, Y. H., Tai, S., & Meng, H. M. (2017, January). Data Visualization on Global Trends on Cancer Incidence An Application of IBM Watson Analytics. In Proceedings of the 50th Hawaii International Conference on System Sciences.

Wagner, J., & Staff, P. W. (2014). Machine Learning and Predictive Analytics Foster Growth. Predictive Analytics Times.

Wang, B., Chen, C., Li Liu, D., Asseng, S., Yu, Q., & Yang, X. (2015). Effects of climate trends and variability on wheat yield variability in eastern Australia. Climate Research, 64(2), 173-186.

Zhu, W. D. J., Foyle, B., Gagné, D., Gupta, V., Magdalen, J., Mundi, A. S., ... & Triska, M. (2014). IBM Watson Content Analytics: Discovering Actionable Insight from Your Content. IBM Redbooks.