Some twenty two years back from now, in the year 1988, he worked on a theory for interpreting radon [a gaseous daughter product of uranium mineral] data. It was about using a combination of both soil radium concentration and radon concentration in soil-gas.
As a standard procedure radon concentration in soil sampling holes were contoured to obtain information about the anomalies and its likely relationship with the uranium deposit. The idea he attempted was something like this : Radon is locally produced from soil radium. Thus less radium in soil (xi concentration of radium) means less radon (yi : concentration of radon) and more radium mean more radon in sampling holes which in other words means radon and radium profile should be well correlated (represented by the linear relation yi = mxi + c). In case there is contribution of long distance migrated radon in that area due to buried uranium deposit then mismatch should occur and correlation should be poor.
If the surveyed soil area responds to long distance migration of radon then it is found to influence the intercept parameter c under varying amounts of transported radon contribution to the surveyed area. Both m (slope) and c are influenced but c is proved to be a better parameter. This analysis he presented in a paper entitled “Statistical model for the recognition of a buried uranium deposit” in 1989, volume 27, page 128 – 131 of the ‘Indian journal of pure and applied physics’. After its publication in March, in the same year there was a media headline : The Times of India, Bombay wrote with the headline “New method to find uranium”. In that headline PTI (Press Trust of India) wrote on October 16 : “A new method to detect underground uranium deposits has been developed by Dr.B.K.Bhaumik of the Atomic Minerals Division in Hyderabad. The technique based on soil measurement of radon gas, a byproduct of radioactivity of uranium, will help discovery of deposits of this nuclear fuel. Presence of radon as an indicator of uranium has been used for some 20 years, but scientists had no idea of knowing whether it is due to the presence of uranium in the soil itself or due to a buried deposit nearby. Dr.Bhaumik has now developed a mathematical linear regression model with which soil radon concentration data can be analyzed to identify whether or not the radon emission is due to a buried source. The usefulness of this technique has been confirmed in field trials in the Singhbhum area, according to Dr.Bhaumik’s report published in the Indian Journal of Pure and Applied Physics”.
In the year of 2008 he came across a research paper titled “Radon occurrence in soil – gas and groundwater around an active landslide” in ‘Radiation Measurements’ journal No.43, page 98 – 101 authored by R.C. Ramola, V.M.Choubey, M.S.Negi, Yogesh Prasad and Ganesh Prasad. In this work on page 100 they indicated an observation that “no correlation was observed between radon exhalation rate and uranium and thorium contents in the soil”. This was followed by the sentence “resulting radon concentration in the area is due to migration of radon from deeper part of the earth’s crust --- ”.
Poor correlation between radon concentration and radium profile of soil was proved to be an interesting phenomenon to indicate a different source of radon other than the soil in the immediate vicinity.
In 1989, statistical significance score [research paper : statistical significance analysis ---- ], calculated using airborne radiometric U, U/Th and U/K data, in association with Shri Bhisma Kumar, highlighted an area lying within 79 deg 0 min - 79 deg 15 min long. range and 16 deg 30 min - 16 deg 45 min lat. range, having very high uranium potential proved three years later to be exactly Lambapur Uranium Deposit.
For processing multivariate geological data, it is required that the variables (e.g., concentration of Transition elements, Rare Earth elements over a suite of samples) be normally distributed. Generalized power transform [Box-Cox] is suitable for this purpose where parameter (lambda) is estimated using objective function involving skewness and kurtosis of transformed data. A novel method is proposed based on the difference between observed distribution of frequencies and theoretical normal frequencies [1995 paper : defining pseudo chi-square ---- and 1997 paper : Improved estimation of Box-Cox transformation ----- ) which is found to work well.
In 2010, on Thursday, April 15, The Times of India, New Delhi edition had a report with the Headline "Not told about risk, traders grope in dark" - it was about radioactive radiation leak in Mayapuri scrap market, New Delhi. As reader goes through the lines, there is a box which reads : According to the police, the fresh radiation was detected by a team of experts led by Dr Bhowmick of the emergency response cell of BARC located at R K Puram.