BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:Asia/Kolkata
X-WR-TIMEZONE:Asia/Kolkata
BEGIN:VEVENT
UID:83@cds.iisc.ac.in
DTSTART;TZID=Asia/Kolkata:20241127T093000
DTEND;TZID=Asia/Kolkata:20241127T103000
DTSTAMP:20241114T143831Z
URL:https://cds.iisc.ac.in/events/ph-d-thesis-colloquium-cds-investigation
 -of-the-indian-summer-monsoon-employing-statistical-and-machine-learning-m
 ethods/
SUMMARY:Ph.D. Thesis {Colloquium}: CDS: "Investigation of the Indian summer
  monsoon employing statistical and machine learning methods"
DESCRIPTION:DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES\nPh.D. Thesis Col
 loquium\n============================================================\nSpe
 aker : Ms. AKANKSHA MANOJ RAJAK\nS.R. Number : 06-18-01-10-12-18-2-16460\n
 Title : "Investigation of the Indian summer monsoon employing statistical 
 and machine learning methods"\nResearch Supervisor : Dr. Deepak Narayanan 
 Subramani\nDate &amp\; Time : November 27\, 2024 (Wednesday)\, 9:30 AM\nVe
 nue : # 102 CDS Seminar Hall\n============================================
 =================\nABSTRACT\nThe Indian summer monsoon rain contributes ap
 proximately 70-90 % of India's annual precipitation\, profoundly impacting
  agricultural productivity\, water resource management\, and thereby the I
 ndian economy. The summer monsoon makes onset over the coast of Kerala aro
 und the first week of June and is characterised by spells of more than ave
 rage rain (called active spells) and less than average rain (called as bre
 ak spells). This thesis presents an analysis of the detection and forecast
 ing of the onset date of the Indian summer monsoon using change point dete
 ction methods and deep neural network models. We also study the utility of
  unsupervised statistical methods in detecting active and break spells.\n\
 nIn the first part of the thesis\, statistical change point detection meth
 ods are applied to a 114-year gridded dataset of rainfall in India to dete
 ct monsoon onset\, and active and break phases of the monsoon in an unsupe
 rvised manner. We apply and study the results from multiple Change Point D
 etection methods such as the Pruned Exact Linear Time (PELT) method for ef
 ficient sequential change detection\, Bayesian Online Change Point Detecti
 on (BOCD) for real-time onset detection\, and Topological Data Analysis (T
 DA) for pattern recognition in complex data structures. These methods succ
 essfully identify monsoon onset dates from rainfall time series and establ
 ish relationships between the detected change points and their interannual
  variability. Furthermore\, CPD methods systematically identify active and
  break spells in monsoon patterns by analyzing statistical property shifts
  between change points across multiple locations in India.\n\nIn the secon
 d part of the thesis\, spatiotemporal deep neural network models were deve
 loped to forecast the onset date with a 30- to 60- day lead time. Reanalys
 is and remotely sensed atmospheric and oceanic data including sea surface 
 temperature (SST)\, total cloud cover (TCC)\, net top thermal radiation fo
 r the clear sky atmosphere (TTRC) or outgoing longwave radiation (OLR) and
  mean sea level pressure (MSLP) from 1 March to 30 April are used as input
  to the deep neural models. Data were obtained from the ERA5 reanalysis\, 
 the MODIS Aqua and Terra satellites\, and the NOAA satellites. Three diffe
 rent neural architectures were developed to predict the onset of monsoon f
 rom the above data: (i) ConvLSTM\, (ii) Depthwise Separable Convolution\, 
 and (iii) Transformers. For the first two architectures\, the daily spatia
 l data of the input variables were fed into the model in a full space repr
 esentation and a reduced space representation using convolutional auto-enc
 oders. For the third architecture\, only the reduced space representation 
 was used. Separate models for each input variable\, combined models for al
 l input variables\, and models with different lead times were developed\, 
 and the results were noted. For each model\, extensive ablation studies we
 re conducted. Impact of differe= nt data mix (reanalysis\, remotely sensed
 ) used for training is quantified. Our analysis revealed that SST and TTRC
  can provide accurate predictions with an error margin of less than 3 days
  by the end of March (60-day lead time)\, while variables like Cloud Fract
 ion and MSLP require data from April (30-day lead time) to contribute effe
 ctively to the onset prediction. The best performing neural models for eac
 h variable are stacked together using a meta learner and the final onset d
 ate is predicted. This model achieves a mean absolute error of 1.8 days wi
 th a correlation coefficient of 0.89 at a 30-day lead time and 2.0 days wi
 th a correlation of 0.87 at a 60-day lead time. The existing dynamical and
  data driven models have a 4-day error\, and we show supe rior predictive 
 ability.\n\nThe enhanced prediction capabilities of the developed model pr
 ovide valuable lead time for agricultural planning and water resource mana
 gement\, potentially improving decision-making processes for marginal farm
 ers\, crop-insurance providers and government irrigation departments acros
 s India.\n\n==============================================================
 ==\nALL ARE WELCOME
CATEGORIES:Events,Ph.D. Thesis Colloquium
END:VEVENT
BEGIN:VTIMEZONE
TZID:Asia/Kolkata
X-LIC-LOCATION:Asia/Kolkata
BEGIN:STANDARD
DTSTART:20231128T093000
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR