Accurate and reliable climate data is critical for assessing the risk of climate change to our society's well-being. Increases in temperature, sea-level, and extreme weather events can render many aspects of our society vulnerable including our health, natural resources, and energy-systems (Nicholls & Cazenave, 2010;Trenberth, 2012). Local and regional climate future projection data is the most crucial for planning and mitigating these risks, but is often the least reliable (Schmidt, 2010). Currently, Earth System Models (ESMs) used for simulating Earth's past climate and future projections are often computationally limited to coarse horizontal resolutions, generally between 1° and 3° (Vandal et al., 2017). These low resolution models fail to accurately simulate important physical processes such as precipitation extremes (Kharin et al., 2007). Recent advances in computing resources have allowed for global ESMs to be run at higher resolutions (∼0.25°) for longer time periods and have been shown to improve the simulations of regional mean climate as well as extremes (Mahajan et al., 2015;Wehner et al., 2010). However, these high resolution models remain prohibitively expensive.A computationally less expensive approach to derive high resolution climate data over a region of interest is to map data from low resolution global model simulations to high resolution grids using dynamic or statistical downscaling techniques. Dynamical downscaling involves running high resolution regional dynamical models to extrapolate large scale boundary conditions obtained from a coarser global ESM to finer resolutions on regional scales. Statistical downscaling (SD) aims to map coarse resolution data to high resolution projections using statistical methods like linear regression. Recent studies have shown that machine learning techniques, like neural networks (Fistikoglu & Okkan, 2011;Vu et al., 2016) and support vector machines (Ghosh, 2010), for SD significantly outperform other traditional SD methods. In this study, we use one such computer vision approach called super-resolution (SR), which generates a high resolution image from its low resolution equivalent. SR techniques attempt to generalize across images and have been shown to learn local scale patterns more efficiently than other downscaling methods (Vandal et al., 2017).

Drugs of abuse are known to alter activity in areas of brain associated with reward, cognition and decision making. Changes in neural activity in these regions that follow repeated exposures to abused substances may underlie the development of addictive behaviors and contribute to the high rates of relapse associated with drug use. Measuring real-time changes in neural activity during drug seeking and taking is important for correlating changes in behavior with alterations in neuronal signaling typically measured using ex vivo electrophysiological recordings. In this study, C57BL/6J mice or Sprague-Dawley rats were injected in different brain areas with adeno-associated viruses (AAV) encoding the calcium sensor GCaMP6f along with an optical fiber. Calcium-dependent fluorescence was monitored in the nucleus accumbens core or mPFC during and following exposure to toluene vapor and in the medial prefrontal cortex (mPFC) and orbitofrontal cortex (OFC) during ethanol drinking. Toluene vapor, at concentrations previously shown to induce conditioned place preference, produced a rapid decrease in the frequency of calcium transients in the NAc core of rats that recovered following washout of the toluene vapor. In a probabilistic risk task, GCaMP6 signals in rat mPFC increased just prior to lever pressing and showed decreases during the reward phase that were proportional to reward size. Toluene pretreatment elevated the signal during the decision-making period while post-lever responses were independent of reward size. Using the drinking in the dark (DID) protocol in mice, we observed a consistent increase in GCaMP6 fluorescence during the period leading up to an ethanol drinking bout, a decrease during consumption and a rebound increase following the bout. The initial increase in signal prior to consumption was greater for ethanol and sucrose than water.GCaMP6 signals in the lateral OFC also decreased during ethanol consumption and increased following bout completion while no increase in activity was noted prior to bout initiation. Following repeated cycles of chronic intermittent ethanol (CIE) exposure that enhanced ethanol consumption, OFC calcium signals during and after ethanol drinking were similar to those in air-treated animals.Addition of quinine to the ethanol solution augmented the decrease in signal during consumption in both air and CIE mice while having no effect on the magnitude of the rebound in activity. Conversely, when sucrose was added to the ethanol solution, air exposed mice showed blunted changes in GCaMP6 signals while those in CIE mice were enhanced. Overall, the results from these experiments complement and extend data from prior behavioral and electrophysiological studies and support the use of in vivo fiber photometry in the study of effects of abused substances on brain function.

We present a first application of a fast super resolution convolutional neural network (FSRCNN) approach for downscaling climate simulations. Unlike other SR approaches, FSRCNN uses the same input feature dimensions as the low resolution input. This allows it to have smaller convolution layers, avoiding over-smoothing, and reduced computational costs. We further adapt FSRCNN to feature additional convolution layers after the deconvolution layer, we term FSRCNN-ESM. We use highresolution (0.25°) monthly averaged model output of five surface variables over a part of North America from the US Department of Energy's Energy Exascale Earth System Model's control simulation. These high-resolution and corresponding coarsened lowresolution (1°) pairs of images are used to train the FSRCNN-ESM and evaluate its use as a downscaling approach. We find that FSRCNN-ESM outperforms FSRCNN and other methods in reconstructing high resolution images producing finer spatial scale features with better accuracy for surface temperature, surface radiative fluxes and precipitation.

No abstract

We construct a novel Multi-Input Multi-Output Autoencoder-decoder (MIMO-AE) to capture the non-linear relationship of Southern California precipitation and tropical Pacific Ocean sea surface temperature. The MIMO-AE is trained on both monthly TP-SST and SC-PRECIP anomalies simultaneously. The co-variability of the two fields in the MIMO-AE shared nonlinear latent space can be condensed into an index, termed the MIMO-AE index. We use a transfer learning approach to train a MIMO-AE on the combined dataset of 100 years of output from a historical simulation with the Energy Exascale Earth Systems Model version 1 and a segment of observational data. We further use Long Short-Term Memory networks to assess sub-seasonal predictability of SC-PRECIP using the MIMO-AE index. We find that the MIMO-AE index provides enhanced predictability of SC-PRECIP for a lead-time of up-to four months as compared to Niño 3.4 index and the El Niño Southern Oscillation Longitudinal Index.

The diversity of the spatial pattern of El Niño Southern Oscillation (ENSO) events has limited our understanding of ENSO-associated predictability of regional water cycles. Various indices have been used to characterize and quantify the strength of unique facets of ENSO. A recent study (Williams & Patricola, 2018) proposed a unified approach to characterize ENSO's spatial diversity with a single non-linear index, namely the ENSO longitudinal index (ELI). It is calculated as the average longitude over the tropical Pacific where the sea surface temperature (SST) is above the threshold for deep convection (Williams & Patricola, 2018). Consequently, ELI represents the location of deep convection and the upwards branch of the Walker circulation over the tropical Pacific and tracks their zonal shifts associated with ENSO. These movements directly impact ENSO teleconnections to the mid-latitudes by modulating the extra-tropical wave-trains that impact moisture transport, storm track activity, etc. over remote regions (Patricola et al., 2020). ELI thus has been shown to be more effective at capturing teleconnections to seasonal mean and extreme precipitation regions like California and Southeastern US as compared to other conventional fixed domain indices like Niño 3.4 index (Patricola et al., 2020;Williams & Patricola, 2018).The simulation of regional precipitation and its extremes remains a challenge for Earth System Models (ESMs). Higher resolution ESMs resolve more fine scale features than prevalent low resolution (100 km in the atmosphere) models representing more realistic orographic lifting, vertical mass fluxes, coastal processes, land use as well as mesoscale ocean eddies, although still relying on parameterization for sub-grid scale processes like convection. High-resolution (HR) global models generally appear to improve the simulation of mean and extreme precipitation as compared to their low resolution counterparts, producing more intense precipitation, which can

No abstract

Transformer-based models have demonstrated much success in various natural language processing (NLP) tasks. However, they are often vulnerable to adversarial attacks, such as data poisoning, that can intentionally fool the model into generating incorrect results. In this paper, we present a novel, compound variant of a data poisoning attack on a transformer-based model that maximizes the poisoning effect while minimizing the scope of poisoning. We do so by combining the established data poisoning technique (label flipping) with a novel adversarial artifact selection and insertion technique aimed at minimizing detectability and the scope of the poisoning footprint. We find that using a combination of these two techniques, we achieve a state-of-the-art attack success rate (ASR) of ~90% while poisoning only 0.5% of the original training set, thus minimizing the scope and detectability of the poisoning action.