Abhishek Anand

I am a postdoctoral research scientist in Westervelt Aerosol Lab in Lamont-Doherty Earth Observatory at Columbia University. I am building machine/deep learning models for estimating air pollutant using large datasets from NASA's and European satellite fleet, validated with ground-level sensors in sub-Saharan Africa. I am further using the datasets to investigate health impacts to aid effective local policymaking.

I graduated with Ph.D. in Mechanical Engineering from Carnegie Mellon University. Advised by Prof. Albert Presto, I worked on developing low-cost techniques to measure atmospheric particulate matter and using air pollution data from low-cost monitors to identify emission sources. Additionally, I built a deep learning-based forecast model for fine particulate matter over Pittsburgh (USA) by using novel features from NASA's GEOS-CF and ground level measurements from a low-cost sensor network, spatially deployed over Pittsburgh in Pennsylvania.

I earned my M.Phil. and M.Sc. from Hong Kong University of Science and Technology and B.Tech. in Civil Engineering from Indian Institute of Technology Delhi in New Delhi, India.

Email  /  CV  /  Google Scholar  /  Linkedin; My Columbia Wesbite

Research Interests

I'm interested in application of image processing and machine learning on remote sensing datasets, atmospheric simulations, and investigating health impacts from air pollution.

Publications

(* denotes corresponding authors)

In this paper, we built deep learning model to predict fine scale NOx from traffic at street-level using data from mobile air quality sensors on buses and crowd-sourced Google real-time traffic status.

In this paper, we used the image processing method from Anand et al. (2023) to measure black carbon in African cities using BAM tapes.

In this paper, we deployed computer vision to build an image processing method to estimate atmospheric black carbon from particulate deposits on filter tapes.

In this paper, we deployed wavelet analysis and lowest percentile methods to quantify contributions of traffic-related emissions to on-road gaseous and particulate levels.

In this paper, we developed an innovative solution for remotely measuring sulfur content in ship fuels from stack emissions with low-cost sensors attached on a drone. This work is aimed at effective policy implementation to reduce SO2 emissions from ships.

In this paper, we propose a novel Temperature Look-Up (TLU) model for NO2 gas sensor outputs in case of long-term application, showing an improved performance over existing ML and MLR methods.