Methane hydrates are solid clathrate hydrate compounds where large amounts of methane are enclosed within a crystal structure of water. This crystal is a solid, reminiscent of ice. It can be found in certain areas in the seafloor and in some Arctic permafrost regions. Just like methane, methane hydrates also burn when you put fire to them. If methane hydrate deposits melt, methane is released into the atmosphere, which can contribute to CO2 emissions and global warming. It is therefore of paramount importance to know where such deposits are situated to prevent emissions, as well as enabling potential harnessing of this valuable energy source.
Now (2021), scientists at Sandia National Laboratories have designed an innovative system, which they tested off the coast of North Carolina, to determine with what probability methane hydrates occur in the seafloor. This system comprised probabilistic modelling and machine-learning algorithms. For testing the system, the area around Blake Ridge, a hill on the seafloor southeast of North Carolina’s Outer Banks was chosen, where a known deposit of methane hydrates is located. The Naval Research Laboratory was responsible for providing specific information on the area, such as temperature, total carbon concentration and pressure. Missing data was completed by the advanced machine-learning algorithms that could calculate the missing value on the basis of information about other geologically-similar areas. The scientists at Sandia imported the data from the Naval Research Laboratory into their own original software, which was designed for statistical sampling and analysis. This helped them to calculate the most likely value for influential seafloor properties, as well as the natural variation for the values. Then, using a statistical approach, they inserted a value from the expected range for each property into another software program they had designed, which could model how chemicals reacted and materials moved underground or under the seafloor. The team conducted several methane production simulations of the Blake Ridge region. They also made the discovery that methane gas formed closer to shore. Their software will be freely available for other researchers to use.
So far, the system has successfully been used to create models of a region of the Norwegian Sea between Greenland and Norway and the shallows of the Arctic Ocean offshore of the North Slope of Alaska. The team has also created a much less detailed model of the whole globe and started analysing the mid-Atlantic, where methane was observed leaking from the seafloor a few years ago.
Methane hydrates have been at the centre of scientific attention for many years. In 2017, scientists used published experimental data, ranging from 1940 to 2016, for modelling the initial stability conditions of methane hydrate in pure water with the help of a Classification and Regression Tree (CART). They chose least squares support vector machine (LSSVM) and artificial neural network (ANN) methods as a basis for comparison. Finally, the Leverage mathematical approach was used for evaluating the quality of the data as well as the correctness of the CART model. The study showed that the developed tree-based model gave excellent and accurate results. In addition, applying the Leverage algorithm demonstrated that the CART model was statistically valid and correct, even if the experimental data published in the literature were of varying quality.
In 2019 scientists developed and validated a clathrate hydrate model for gas replacement. They created an online parameter identification technique intended for automatic tuning of model parameters in the field. This model was employed to analyse the behaviour of the laboratory-scale data for methane hydrate formation and decomposition. Then, the model was validated with the field data of the Prudhoe Bay Unit on the Alaska North Slope during 2011 and 2012. In this field experiment, mixed CO2 (i.e., CO2 + N2) was used as a replacement agent for natural gas recovery. The outcome of the experiment showed that the proposed formulation had a good performance with a maximum absolute average relative deviation (AARD) of about 2.83% for CH4, 0.84% for CO2 and 1.67% for N2.
Assessing the global amount and behaviour of methane hydrates holds several benefits: from an ecological point of view, it might be sensible to assess the possible dangers arising from such deposits in view of constantly rising global temperatures; also, these deposits could be harnessed as a source of natural gas for energy production; if leaks can be predicted reliably, possibly negative impacts on the ocean ecology and nutrient cycles can be averted and then harnessed successfully and sustainably.
For now, scientists need to carry out thorough testing on how effective this model actually is in predicting methane seeps on the seafloor. They also want to see if they can estimate the distribution of methane seeps and whether they correspond to the thermodynamic properties of methane-hydrate stability. The project will receive further funding from the Naval Research Laboratory.