Search

Astronomers require efficient automated detection and classification pipelines when conducting large-scale surveys of the (optical) sky for variable and transient sources. Such pipelines are fundamentally important, as they permit rapid follow-up and analysis of those detections most likely to be of scientific value. We therefore present a deep learning pipeline based on the convolutional neural network architecture called MeerCRAB. It is designed to filter out the so called "bogus" detections from true astrophysical sources in the transient detection pipeline of the MeerLICHT telescope. Optical candidates are described using a variety of 2D images and numerical features extracted from those images. The relationship between the input images and the target classes is unclear, since the ground truth is poorly defined and often the subject of debate. This makes it difficult to determine which source of information should be used to train a classification algorithm. We therefore used two methods for labelling our data (i) thresholding and (ii) latent class model approaches. We deployed variants of MeerCRAB that employed different network architectures trained using different combinations of input images and training set choices, based on classification labels provided by volunteers. The deepest network worked best with an accuracy of 99.5% and Matthews correlation coefficient (MCC) value of 0.989. The best model was integrated to the MeerLICHT transient vetting pipeline, enabling the accurate and efficient classification of detected transients that allows researchers to select the most promising candidates for their research goals. [ABSTRACT FROM AUTHOR]

In the absence of environmental steering, tropical cyclone (TC) motion largely reflects "beta drift" owing to differential planetary vorticity advection by the storm's outer circulation. It is known that model physics choices (especially those relating to convection) can significantly alter these outer winds and thus the storm track. Here, semi-idealized simulations are used to explore the influence of the initialization on subsequent vortex evolution and motion. Specifically, TCs bred from a buoyant "bubble" are compared to bogussed vortices having a wide variety of parameterized shapes and sizes matching observations. As expected, the bogussed storms commencing with the strongest outer winds propagated fastest and, as a result, huge structure-dependent position differences quickly appeared. However, the forward speed variation among the initially bogussed TCs subsequently declined as a progressive homogenization harmonized the initially supplied structural differences. The homogenization likely involved model physics such as microphysics. This result casts doubt on the ability of models to retain and propagate forward information supplied at the initialization by advanced data assimilation techniques or parameterized vortex wind profiles. Asymmetries in near-core convective heating emerged as an important structural aspect that survived the homogenization tendency. The bubble and bogussed TCs developed markedly different heating patterns, which appear to help explain why the artificially-established storms tended to move about three times faster than their bubble counterparts. The reasons for this are not presently understood fully.

In the absence of environmental steering, tropical cyclone (TC) motion largely reflects "beta drift" owing to differential planetary vorticity advection by the storm's outer circulation. It is known that model physics choices (especially those relating to convection) can significantly alter these outer winds and thus the storm track. Here, semi-idealized simulations are used to explore the influence of the initialization on subsequent vortex evolution and motion. Specifically, TCs bred from a buoyant "bubble" are compared to bogussed vortices having a wide variety of parameterized shapes and sizes matching observations. As expected, the bogussed storms commencing with the strongest outer winds propagated fastest and, as a result, huge structure-dependent position differences quickly appeared. However, the forward speed variation among the initially bogussed TCs subsequently declined as a progressive homogenization harmonized the initially supplied structural differences. The homogenization likely involved model physics such as microphysics. This result casts doubt on the ability of models to retain and propagate forward information supplied at the initialization by advanced data assimilation techniques or parameterized vortex wind profiles. Asymmetries in near-core convective heating emerged as an important structural aspect that survived the homogenization tendency. The bubble and bogussed TCs developed markedly different heating patterns, which appear to help explain why the artificially-established storms tended to move about three times faster than their bubble counterparts. The reasons for this are not presently understood fully. [ABSTRACT FROM AUTHOR]