Title:
Modeling the effects of weather conditions on agriculture using spatial data mining theories
Abstract
Weather forecasting is one of the most important scientific and technological issues and challenges of the century problems. Extreme weather conditions can kill plants and damage whole farms which may result in huge agricultural loss.
In this thesis, necessary weather data including precipitation, temperature, wind speed, solar radiation, and humidity will be collected from meteorological sites. GIS will be used to create a geo-database in order to connect the data to their location and take into account the geographic effects. Adaptive and dynamic spatial data mining will be applied to forecast weather conditions and alert farmers ahead of time to minimize agricultural losses. Cross validation will be implemented to test the results and insure satisfying modeling to the phenomena. The main aim is to be able to predict future disasters and suspicious locations to suggest the right handling and try to stop or minimize the suspected damage. All the results will be presented in thematic digital maps to highlights the spatial analysis results and effect.
Introduction
Agriculture plays a critical role in the entire life of the economy of a given country as it provides food, raw material, and employment opportunities to many people.
People working in agriculture in the developing countries are typically much poorer than people who work in other sectors due to loss of crops as a result of abnormal natural disasters (snow and frost, hurricanes, floods, fires and tornadoes).
Weather conditions including precipitation, temperature, wind speed, solar radiation, and humidity affect plants in different ways that can make them more susceptible to disease and insect problems. Unlike short term weather, climate represents the average weather conditions over a long period of time which determines what will probably grow well in a certain region. Extreme weather conditions can kill plants and damage whole farms which may result in huge agricultural loss.
Meteorological data mining can be used to find hidden patterns within the available meteorological data and to retrieve information leading into useful knowledge. Useful knowledge can play an important role in understanding the climate variability and climate prediction. This understanding can be used to support many important sectors such as agriculture, plants, and water resources [1].
The anticipated outcome of this research is to help farmers reducing their risks by protecting their farms from diseases and other agricultural dangers in an attempt to increase their incomes. In this research, necessary weather data including precipitation, temperature, wind speed, solar radiation, and humidity will be collected from meteorological sites and apply adaptive and dynamic data mining methods to forecast weather conditions and alert farmers ahead of time to minimize agricultural losses. Sources of raw data will be collected from the Palestinian Meteorological department, the Palestinian ministry of agriculture, and from online sources.
Such weather data can be used to predict plants diseases and insect problems for farmers so that they can take precautions and safety tools to protect their farms from diseases or death.
Problem statement
Weather forecasting for the future is one of the most important attributes and variables in an agriculture sector. Variability of weather and climatic factors combined with location and geographic effects, especially those atmospheric parameters will be the major attributes for make future prediction for agriculture disasters. When we think about the food we eat on daily basis, we have to think about the importance of the success of agriculture that could be achieved by protecting farms from natural disasters. Many farmers lose their farms because of some diseases related to the effect of unexpected weather changes. Extreme weather conditions can kill plants and damage whole farms which may result in huge agricultural loss. This thesis proposes a new idea of implementing spatial data mining theories for predicting the agricultural dangers depending on data collected from meteorological site in order to protect farmers, their farms, and agriculture.
Motivation
There are many countries that rely heavily on agriculture in the national income, while agriculture occupies a special place in many developed countries. Thus, any effort that helps to reduce the exposure to the weather or is likely to lead to significant global benefits, both economic and social climate risks.
In particular, while the large amount of information on weather and climate is now available to farmers and some types of information under development or already running, particularly climate prediction; it may not be suitable for use by the farmers to take their decision. Studies indicate that agricultural decisions that are the result of weather and climate information that is clear and well-defined. So that weather and climate information can be formed to fit well agricultural decisions. There are many weather web sites and many TV channels that broadcast weather on daily basis. But those sources do not give the dangers could be happening due to weather and climate information location.
This research will use data mining theories and geographic information to predict the agricultural dangers spatially depending on data collected from meteorological sites, these predictions will help farmers to take precautions and safety to protect their farms from damage.
Contributions
The contribution of our research is to model the effect of the weather and the geographic parameters on the success in the agriculture field.
The proposed intelligent model will help to predict severe weather conditions and alert farmers to take precautions to avoid agricultural damage taking into account location and time.
Thesis scope
��� Programming collection and storing the data from meteorological sites
��� Building Geo-database using GIS environment and tools. Analyze collected data based on spatial data mining theories.
��� Modeling and prediction the spatial distribution of weather parameters and their effect on agriculture.
��� Using the results for making and supporting decisions in the agriculture field.
Research methodology
A program with an intelligent model will be developed to study the effect of weather conditions and other geographic parameters on agriculture based on available meteorological and agricultural data. That model will be able to parse meteorological sites and collect data about current weather and forecast the future weather conditions.
The program must store all data at our database with a format suitable for analysis. That database will be in the form of relational database as well as GIS (Geographic Information Systems) database that stores, presents and analyzes spatial data.
Data preparation is the next step for analysis using spatial data mining theories to be able to study and predict any agricultural dangers that could be happen as a result of extreme weather conditions.
Using the GIS as Decision support system to implement multi dimension analysis, visualizing and modeling help minimize the danger and damage in the agriculture domain taking into account .
Related works:
Useful knowledge can play an important role in understanding the climate variability and climate prediction. The knowledge discovery process comprises six phases: data selection, data cleaning, enrichment, data transformation, data mining, and the reporting and display of the discovered information [2].
There are many factors affecting the weather, which are directly related to animal and vegetable farms, such as humidity, temperature and wind speed. Laila Mohamed and her colleagues show the number of affected animals for each disease when the temperature is hot and humidity is low. They used a data mining technique to develop a system that can be used to analysis and measures the effect of climate on animal production. This system first combines the BOVIS database with the weather database using the data warehouses techniques. The analysis results of the discovered knowledge show that the system can be used to predict the disease occurrence. The future work will be extended using more attributes from the unused variables in both BOVIS and weather database like, animal feed and rain. And it can be integrated with other data mining algorithms to generate new types of discovered knowledge [3]. Meteorological databases from the last years contain enough data in the farms production business to allow the modeling of farms losses due to the weather. Vale and his colleges analyze databases of poultry production associated to climate data using data mining. Attribute selection, data classification and decision trees were used to model and predict the effect of heat wave incidence on broiler mortality. The authors showed that proposed models depend on the available data. Meteorological stations generate a large number of data that are seldom used for animal production or agriculture Thus, the development of models using regional or local data to predict production losses may be useful for producers [4].
Meteorological data mining is a form of data mining concerned to find hidden patterns within the meteorological data available to a large extent, so that it can be retrieved information is transformed into usable knowledge. Sara N and her college extract useful knowledge from daily historical weather data collected locally at Gaza Strip city. The data include nine years period [1977-1985]. After data preprocessing, they apply outlier analysis, clustering, prediction, classification and association rules mining techniques. After each mining technique, they present the extracted knowledge and describe its importance in meteorological field. They propose building adaptive and dynamic data mining methods that can learn dynamically to match the nature of rapidly changeable weather nature and sudden events as a future work [5].To make data mining algorithms more accurate in predicting it���s important to store the meteorological data where it becomes cumulative over time, in addition to that it���s important for taking the new data, which can occur suddenly. Disasters Saptarsi Goswami and his college focus on reviewing the application of data mining and analytical techniques designed so far for (i) prediction, (ii) detection, and (iii) development of appropriate disaster management strategy based on the collected data from disasters. A detailed description of availability of data from geological observatories (seismological, hydrological), satellites, remote sensing and newer sources like social networking sites as twitter is presented. They propose framework for building a disaster management database for India hosted on open source Big Data platform like Hadoop. They observe in their study that, there have not been enough works done in this area to tap the potential of these sources especially in the context of India. They propose to build a data store for natural disasters from these sources in Phase 1. In Phase II, we intend to integrate it with other sources of information [6]. Also Folorunsho Olaiya and his college investigate the use of data mining techniques in forecasting maximum temperature, rainfall, evaporation and wind speed. The results show that given enough data the observed trend over time could be studied and important deviations which show changes in climatic patterns identified. This was carried out using Artificial Neural Network and Decision Tree algorithms and meteorological data collected between 2000 and 2009 from the city of Ibadan, Nigeria. In future research works neuro-fuzzy models will be used for the weather prediction process. This work is important to climatic change studies because the variation in weather conditions in term of temperature, rainfall and wind speed can be studied using these data mining techniques [7].
Our suggested work will take advantages of other related works; we propose to develop windows application that will collect data for all attributes from metrological sites that will help us in forecasting the weather. Also our data will be updated automatically, and then data mining methods will be used to predict sudden weather changes.
Steps of the thesis
1. Develop program (widows application)
2. Building geo-database and GIS format data.
3. Spatial Analysis on the collected data based on data mining theories.
4. Validation, verification, and calibration of models
5. Repeat 3 and 4 until we reach sufficiently accurate simulators
Expected results and planning
1- Build new windows application that will collect information about weather.
2- Tune simulator to be suitable for our data
3- Support decision system to enhance prediction of weather dangers and agriculture handling.