Waste Management has always been of critical concern in places where the public gather the most like in malls, theatres, in academic institutions, industries, residential colonies etc. Even though there are advanced waste sorting and disposal techniques, it has always been a concern for the public residing around the dump yard area about the harmful gases that are emitted from these centers. This paper presents a novel approach to support the waste management centers to monitor and predict harmful gas emissions using a combination of hardware sensors, digital twin technology, and a Python-based web application. The waste disposal area is assumed to be the physical environment. MQ2 and MQ135 sensors are integrated with an Arduino board to detect hazardous gas levels in the environment, triggering an alert mechanism via a buzzer when safety thresholds are exceeded. The threshold is continuously monitored by the system, named DT-ATMOS, which includes modules for data visualization, real-time monitoring, and also predictive analytics using an ARIMA model. A tuned set of parameters has been used to train the model so as to get minimum MSE during the testing phase. The proposed system focuses on supporting Sustainable Development Goal 11.3,11.6 and 11.7 towards attaining safe, resilient cities.