Analysis of 2018 flood over Kerala using GPM real-time datasets
Extreme rainfall and associated landslides can cause enormous damage. Parts of Kerala and a southern district of Karnataka experienced one of the worst landslides in recent decades. There are views that root cause was the environmental degradation by anthropogenic activities. While it may be true, it is also possible that the rainfall during the period was unprecedented and this caused landslides and floods. Therefore, it is important to closely examine the rainfall time series in areas affected by the flood/landslide to understand the nature of rain events that preceded the floods. Ground based rain gauge network in remote and hilly areas are sparse. Now, there are satellite systems capable of providing rainfall information at high temporal and spatial resolutions. Satellites cover large area as well. Here, I make use of Global Precipitation Measurement (GPM) mission data available since 2016. NASA’s Integrated Multisatellite Retrievals for GPM (IMERG) gives rainfall data with a temporal resolution of 30 min and spatial resolution of 0.1o x 0.1o. GPM is advanced over the ground-based measurement systems as it measures precipitation from space. Ground-based equipment cannot collect measurements in entire regions over the globe and have very limited spatial coverage. To know if 2018 monsoon season was unusual, I am analysing the high resolution hourly rainfall data, for south India for the months of June, July and August in 2013, 2014, 2015, 2016, 2017 and 2018. The GPM data has been analysed using matlab. Results indicate that there is a significant change in the amount of monsoon rains in 2018 compared to other five years. The rainfall, especially in Mid-August of 2018, is found to be extremely higher than in the previous years. Thus, the main reason behind the flood and landslide in 2018 was this unprecedented rainfall. The continuous rainfall combined with simultaneous release of water from several dams resulted in a disastrous flooding in some districts of Kerala. This report aims at analysing the individual events, which led to the enormous flood that cost the life of many people and created a huge loss for the state.
Keywords: precipitation, GPM, IMERG, flood, landslides
|GPM||Global Precipitation Measurment|
|IMERG||Integrated Multi-satellite Retrievals for GPM|
|TRMM||Tropical Rainfall Measuring Mission|
|DPR||Dual frequency Precipitation Radar|
|GMI||GPM Microwave Imager|
|NASA||National Aeronautics and Space Administration|
Extreme precipitation, flood, and landslides are natural disasters that affect human life and the economy badly. The amount of precipitation and frequency of flood has risen alarmingly in recent years. Therefore monitoring and analyzing the variability and trends in rainfall events are crucial to the well being of humans. But obtaining accurate precipitation has always been challenging for scientists. Currently, there are three methods to measure precipitation, (i) rain gauge, (ii) weather radar, and (iii) satellite-based sensors. Rain Gauges provide the most straight forward and accurate precipitation observations at the points of installation but their spatial coverage is very limited over land and installation over oceans is extremely hard. Besides, high wind speeds cause error in precipitation measured by rain gauges. In regard to the weather radar, it can provide the internal structure of storms and real-time high-resolution monitoring over an area of about 150 km in radius around the radar. However, the radar coverage is also very limited in India. Remote sensing is a powerful tool because it offers a wide range of information. Precipitation estimation is done by remote sensing using instruments attached to satellites. Therefore estimation of precipitation on a global basis relies on earth observation satellites as they detect rain, snow and fog and can monitor weather in real-time.
NASA’s (National Aeronautics and Space Administration) GPM (Global Precipitation Measurement) is an international constellation of satellites that offers a new generation of global measurements of the precipitation (rainfall and snow) with high spatial and temporal resolution. It is composed of a main core observatory satellite (GPM CO) and ten other satellites working together to provide a near real-time estimation of the precipitation distribution all over the globe. GPM rainfall estimates have been validated using ground measurements. GPM data are extensively used for understanding precipitation events and for weather forecasts.
The south Indian state of Kerala was affected by several floods in the year 2018. The number of deaths is estimated to be 440 and economic damage of around $3 billion dollars. The unprecedented rainfall at the end of July and mid-August was the major reason, according to many studies. The continuous rainfall for the entire month of August 2018 impacted human life severely and resulted in a huge loss of life and property.
Idukki was the most affected as the heavy rainfall caused landslides in the hilly district. According to reports Idukki received some of the highest rainfall in these two months. Heavy rainfall over the Idukki dam caused the water level to rise at an alarming rate. Idukki dam shutters were opened after 26 years causing a deluge.
In this report we analyze the fluctuations in rainfall over the region of the Idukki dam for the past 6 consecutive years (2013-2018) during the monsoon months of June, July and August using the near-real-time product of IMERGHH (IMERHG half-hourly) datasets.
The region of study is the catchment upstream of the Idukki reservoir that extends up to 650 km2 and lies between 9.7o to 9.9o north and 76.7o to 77.1o east.
Idukki dam is built upon the Periyar River, between the Kuravan and Kurathi hills in Kerala. The reservoir is formed by the Idukki dam, Cheruthoni dam, and the Kulamavu dam. It is the largest reservoir in Kerala with a capacity of 1460 MCM (million cubic meters).
DATA AND METHODOLOGY
Global Precipitation Measurement (GPM) mission is a constellation of satellites, including one Core Observatory satellite and ten partner satellites. It consists of dual-frequency precipitation radar (DPR) and GPM Microwave Imager (GMI) with capabilities to sense light rain and falling snow. It operates on a non-sun-synchronous orbit with an inclination of 65o and allows the GPM CO to measure precipitation from the Arctic to the Antarctic across all hours of the day. The data processing is then done in the Precipitation Processing System (PPS).
The Integrated Multi-satellite Retrievals for GPM (IMERG) is a level 3 NASA product (give reference). It produces half hour global products at 0.1o x 0.1o spatial resolution between 60o north and 60 o south. These are available hours after data collection and can be used for precipitation analysis.
All GPM products are accessible online in the NASA (https://pmm.nasa.gov/data-access/downloads/gpm) as standard HDF5 format files.
Before the GPM, TRMM satellite provided global precipitation data from 1998 to 2015. The IMERG datasets are available online from the month of June 2000. Each IMERGHH data set represents a 30-minute span starting on the hour or half-hour. Thus, the first image of the day includes data for 00:00:000–00:29:59 UTC (Coordinated Universal Time).
The IMERG data has been downloaded in HDF5 format. Any tool that can read standard HDF5 format can be used for processing IMERG files. The grid is a 0.1o x 0.1o global array of points. Individual IMERG grid cells are treated as data points. Its size is 1800 x 3600 with X-axis (latitude) extending from 90o south to 90o north and Y-axis (longitude) from 180o west to 180o east. Each file has nine datasets out of which IR precipitation has been taken. It gives Microwave-calibrated infrared precipitation estimate covering the 30-minute period and the values range from 0 to 1000 mm/hr.
The IMERG data over the Idukki dam reservoir for the months of June, July and August for the past six years (2013-2018) have been downloaded and processed using MATLAB.
The catchment area of Idukki dam is 649.3 km2. Each grid of GMP covers an area of 11 x 11 km. The accumulated precipitation estimates over the dam catchment area have been made considering the coordinate boundaries of the reservoir (76.7o E- 77.1o E, 9.7o N- 9.9o N). The region of study is covered by a two by four (2x4) grid cells each of 0.1o x 0.1o and sums to an area of 968 km2. To calculate the rainfall over the reservoir the value has to be multiplied with a factor of 0.67
(649.3/968=0.67; where 649.3 is the catchment area of the dam.)
We have taken the GPM Half Hourly (GPM HH) data in mm/hr. To calculate the rainfall accumulated in 30 min the value is multiplied by a factor of 0.5.
RHH = G x 0.5; where G is the GPM image grid value.
The daily rainfall is calculated by summing the 48 values of half-hourly data of each day.
To study clearly the Inter Annual Differences in rainfall the daily rainfall has been summed to give a weekly accumulation of rainfall. Bar plots have been used and 3 months are clubbed together in a single plot so as to study the monsoon season of all the years with respect to a common scale. With 7 days in each week and 92 days in the three months, we have plotted 13 (92/7) weeks each year.
Also, graphs of catchment rain volume normalized by dam capacity and cumulative rainfall normalized by dam capacity have been plotted to make the comparison easier.
The following graphs are included in the report.
Graph 1: Bar graphs of weekly rainfall, with 13 weeks each year, for 6 years.
Graph 2: Line graph of cumulative rainfall of the years.
Graph 3: Rainfall volume normalized by dam capacity.
Graph 4: Accumulated catchment rain volume normalized by dam capacity.
Graph 1&2 depicts the weekly observed rainfall distribution and cumulative rainfall of three months respectively.
Graph 1 gives the amount of rainfall in each week over Idukki dam catchment area during the months of June, July and August of 2013-2018.
A wide variation in the rainfall patterns is observed over the years. A peek in rainfall was observed in 2013 with discontinuous heavy downpour and a dip in the subsequent years of 2014 and 2015 resulting in a draught-like situation in 2015. It then began slowly rising in 2016, 2017 and 2018.
Analysis of graph 1 shows that, 2013 had heavy downpour during the 1st, 4th and 7th weeks. The 4th week downpour caused minor floods in some places with no large scale devastations. In the year 2018, the unprecedented heavy rainfall and the non-uniform temporal distribution resulted in severe damage. Scrutiny of data shows that the cumulative rainfall realized during 5th, 6th &7th and 11th week of 2018 were quite significant while the 12th&13th week rainfall was abnormally high.
The first onset of flood occurred towards the end of July as a result of continuous heavy downpour in the 5th, 6th and 7th week with a small scale damage. After the 7th week the rain receded.
The second onset of flood occurred during the middle of august. There were heavy downpours across the state during entire 11th week which increased water levels of various dam reservoirs in the state. Six out of the seven reservoirs had more than 90% of their full reservoir level. Study shows that major reservoirs were almost full before the heavy rainfall in the 11th week. The water resources and dams in the state were prepared for a drought-like situation and managed to store water for irrigation and hydropower generation considering the extremely low amount of rainfall in the previous two years. Therefore, these reservoirs could not accommodate the further flow generated by the heavy downpour and a substantial amount had to be released in the short span of time. Following the dangerously high inflow of water into the catchment area of Idukki dam, it was opened on 9th August, after a gap of 26 years.
Graph 3 shows the rainfall volume in Idukki dam normalized by its capacity. The graph clearly shows that rainfall volume of the 11th and 13th weeks alone were enough to fill the dam by more than half its volume.
Graph 4 shows the accumulated rainfall volume in Idukki dam normalized by its capacity.
While in 2013, the rainfall receded by the starting of July, the rainfall was constantly rising in 2018 and by the end of 11th week the accumulated volume is sufficient enough to fill the reservoir.
Records show that the rainfall in the catchment area of the Idukki dam alone has a return period of more than 100 years. The combination of reservoir water and the unprecedented extreme rainfall worsened the flood situation in Kerala that cost lives of around 450 people across the state.
The intensity of heavy rainfall has varied considerably in the last few years. Kerala received extremely high rainfall in the year 2018 compared to the previous years. The downpour in mid-August had a return period of more than 100 years. Persistent heavy rainfall and the water released from major dams caused severe flood across the state.
Even though Kerala flood can be attributed to the unprecedented heavy rainfall and reservoir storage, climate change also had an important role in it. The increase in extreme precipitation can be attributed to the warming of the planet.
Researchers attributed the extreme rainfall to several factors caused by human activities like the destruction of flood plains and land cover change. Changes in land use methods might have caused the landslides during the floods in the Idukki district. Studies show that global warming has an important role in the increasing trends in extreme events in the last decade like extreme heat events and extreme precipitation. The roles of natural factors like land use and damage to Western Ghats are yet to be examined by scientists,
Considering the frequency of such extreme events climate change on a global scale should be kept in check. Improved forecast of extreme rainfall could help in better management of reservoirs and preparedness for facing such major events to minimise the loss and damage.
· Ali, H. and Mishap, V.: Increase in sub-daily precipitation extremes in India under 1.5 and 2.0°C warming worlds, Geophysics. Res. Lett., doi:10.1029/2018GL078689, 2018.
· Gail Skofronick,Jackson, Dalia Kirschbaum, Walter Petersen, George Huffman :Chris Kidd3,1 Erich Stocker1 Ramesh Kakar4The Global Precipitation Measurement (GPM) mission’s scientific achievements and societal ontributions: reviewing four years of advanced rain and snow observations.
· M. Rajeevan, Jyoti Bhate,and A. K. Jaswal: Analysis of variability and trends of extreme rainfall events over India using 104 years of gridded daily rainfall data.
· Nair, K.M., Chattopadhaya, S. and Sasidharan, C.K. (1997) Analysis of the environmental impact of the lowland to highland migration in the Western Ghats region Kerala, Technical report submitted to the Ministry of Environment and Forests: Thiruvananthapuram, India. CESS, Government of Kerala.
· V.V. Sai Krishna, Anil Kumar Dikshit and Kamal Pandey: Flood modeling with Global precipitation Measurement (GPM) satellite rainfall data – A case study of Dehradun, Uttarakhand, India.
· Central Water Commission, Hydrological Studies Organization: STUDY REPORT KERALA FLOODS OF AUGUST 2018.
·Vimal Mishra, Saran Aadhar, Harsh Shah, Rahul Kumar, D. R. Pattanaik, Amar Deep Tiwari: The Kerala flood of 2018: combined impact of extreme rainfall and reservoir storage.
·Zhaoli Wanga,b, Ruida Zhonga,c, Chengguang Laia,b, Jiachao Chena : Evaluation of the GPM IMERG satellite-based precipitation products and the hydrological utility.
I wish to extend my sincere gratitude to the Indian Academy of Sciences for selecting me under its Summer Student Fellowship Program and my guide Prof. G.S Bhat, Centre for Atmospheric and Oceanic Sciences, Indian Institute of Science, Bengaluru, for giving me an opportunity to carry out this project under his supervision.
I extend my thanks to Dr. Kapil Sindhu and Mr. Gautam Kumar Suman for patiently clarifying my doubts and guiding me in this work. I also thank Prof Jyotirmayee Satapathy for her constant support and guidance.
I would like to thank Prof. Manisha Kumari, Dept. of Earth Science, Pondicherry Central University for recommendation and support and my other teachers in Pondicherry University.
Finally, I sincerely thank my parents and friends for their encouragement and support to complete this project.