A Deep Gaussian Process For Forecasting Crop Yield and Time-Series Analysis of Precipitation
Earlier this year my paper got accepted at IEEE Geoscience and Remote Sensing Symposium hosted by the IEEE Geoscience and Remote Sensing Society which was supposed to be held at Waikolowa, Hawaii, USA in July but due to the coronavirus pandemic it moved to a virtual setting.
The conference was sponsored by NASA - National Aeronautics and Space Administration, Descartes Labs, Quartus Engineering, MDPI, Taylor & Francis, JAXA(宇宙航空研究開発機構) (Japan Aerospace Exploration Agency), Alaska Satellite Facility and will be at the IEEE Xplore digital library.
We replicated a Deep Gaussian process to predict yield of Potato and compared the RMSE of different models. The LSTM network and Time-Series analysis showed how the trends of rainfall have changed over the years. With such analyses, policy makers can make decisions and set policies to help protect our crops and implement risk mitigation strategies to combat extremely high and low precipitation. If not taken care of properly these changes in climate patterns are going to jeopardize Bangladesh's significant achievements in income growth and poverty over the last 20 years. Keeping these in mind, climate change impact assessments must be routinely and systematically compiled, tracked and analyzed.
Did you know:
The agricultural sector has been a crucial source of Bangladesh’s economy and still contributing around 17% of her GDP.
Provides employment to 45% of the labor force.
The rural people’s livelihood mainly depend on agriculture either directly or indirectly and the number is almost 84%.
Global Climate Risk Index 2019 has ranked Bangladesh as 7th most vulnerable to climate risk.
The reported mean annual temperature of Bangladesh has increased during the period of 1895-1980 at 0.310 degree centigrade over the past two decades.
The World Bank stated that more than three quarters or 13.4 crore of around 16.5 crore population of Bangladesh are at the risk of declining living standards as a result of rising temperature and erratic rainfall due to climate change.
It is estimated that an increase of 1 mm/month in rainfall and 10°C of temperature in Bangladesh is expected to lead to a rise of approximately US$ 4-15 for net crops per hectare.
PS: Had to squeeze the video into 5 mins in length because of time constraint.
Paper Link: https://www.researchgate.net/publicat...
Website: https://mostafadidar.com
#climateemergency #climatesolutions #artificialIntelligence #SpaceTech #remoteSensing