Search for: All records

Biomass burning plays an important role in climate-forcing and atmospheric chemistry. The drivers of fire activity over the past two centuries, however, are hotly debated and fueled by poor constraints on the magnitude and trends of preindustrial fire regimes. As a powerful tracer of biomass burning, reconstructions of paleoatmospheric carbon monoxide (CO) can provide valuable information on the evolution of fire activity across the preindustrial to industrial transition. Here too, however, significant disagreements between existing CO records currently allow for opposing fire histories. In this study, we reconstruct a continuous record of Antarctic ice core CO between 1821 and 1995 CE to overlap with direct atmospheric observations. Our record indicates that the Southern Hemisphere CO burden ([CO]) increased by 50% from a preindustrial mixing ratio of ca. 35 ppb to ca. 53 ppb by 1995 CE with more variability than allowed for by state-of-the-art chemistry-climate models, suggesting that historic CO dynamics have been not fully accounted for. Using a 6-troposphere box model, a 40 to 50% decrease in Southern Hemisphere biomass-burning emissions, coincident with unprecedented rates of early 20th century anthropogenic land-use change, is identified as a strong candidate for this mismatch.

 

Response to oxidative stress is universal in almost all organisms and the mitochondrial membrane protein, BbOhmm, negatively affects oxidative stress responses and virulence in the insect fungal pathogen,Beauveria bassiana. Nothing further, however, is known concerning howBbOhmmand this phenomenon is regulated.

Threeoxidativestressresponse regulating Zn₂Cys₆transcription factors (BbOsrR1, 2, and 3) were identified and verified via chromatin immunoprecipitation (ChIP)-qPCR analysis as binding to theBbOhmmpromoter region, with BbOsrR2 showing the strongest binding. Targeted gene knockout ofBbOsrR1orBbOsrR3led to decreasedBbOhmmexpression and consequently increased tolerances to free radical generating compounds (H₂O₂and menadione), whereas the ΔBbOsrR2strain showed increasedBbOhmmexpression with concomitant decreased tolerances to these compounds. RNA and ChIP sequencing analysis revealed that BbOsrR1 directly regulated a wide range of antioxidation and transcription-associated genes, negatively affecting the expression of theBbClp1cyclin andBbOsrR2. BbClp1 was shown to localize to the cell nucleus and negatively mediate oxidative stress responses. BbOsrR2 and BbOsrR3 were shown to feed into the Fus3-MAPK pathway in addition to regulating antioxidation and detoxification genes. Binding motifs for the three transcription factors were found to partially overlap in the promoter region ofBbOhmmand other target genes. Whereas BbOsrR1 appeared to function independently, co-immunoprecipitation revealed complex formation between BbClp1, BbOsrR2, and BbOsrR3, with BbClp1 partially regulating BbOsrR2 phosphorylation.

These findings reveal a regulatory network mediated by BbOsrR1 and the formation of a BbClp1-BbOsrR2-BbOsrR3 complex that orchestrates fungal oxidative stress responses.

 

It remains an interesting and challenging problem to synthesize a vivid and realistic singing face driven by music. In this paper, we present a method for this task with natural motions for the lips, facial expression, head pose, and eyes. Due to the coupling of mixed information for the human voice and backing music in common music audio signals, we design a decouple-and-fuse strategy to tackle the challenge. We first decompose the input music audio into a human voice stream and a backing music stream. Due to the implicit and complicated correlation between the two-stream input signals and the dynamics of the facial expressions, head motions, and eye states, we model their relationship with an attention scheme, where the effects of the two streams are fused seamlessly. Furthermore, to improve the expressivenes of the generated results, we decompose head movement generation in terms of speed and direction, and decompose eye state generation into short-term blinking and long-term eye closing, modeling them separately. We have also built a novel dataset, SingingFace, to support training and evaluation of models for this task, including future work on this topic. Extensive experiments and a user study show that our proposed method is capable of synthesizing vivid singing faces, qualitatively and quantitatively better than the prior state-of-the-art.

 

Estimating fire emissions prior to the satellite era is challenging because observations are limited, leading to large uncertainties in the calculated aerosol climate forcing following the preindustrial era. This challenge further limits the ability of climate models to accurately project future climate change. Here, we reconstruct a gridded dataset of global biomass burning emissions from 1750 to 2010 using inverse analysis that leveraged a global array of 31 ice core records of black carbon deposition fluxes, two different historical emission inventories as a priori estimates, and emission-deposition sensitivities simulated by the atmospheric chemical transport model GEOS-Chem. The reconstructed emissions exhibit greater temporal variabilities which are more consistent with paleoclimate proxies. Our ice core constrained emissions reduced the uncertainties in simulated cloud condensation nuclei and aerosol radiative forcing associated with the discrepancy in preindustrial biomass burning emissions. The derived emissions can also be used in studies of ocean and terrestrial biogeochemistry.

 

Large language models (LLMs) have demonstrated an impressive ability to perform arithmetic and symbolic reasoning tasks, when provided with a few examples at test time ("few-shot prompting"). Much of this success can be attributed to prompting methods such as "chain-of-thought", which employ LLMs for both understanding the problem description by decomposing it into steps, as well as solving each step of the problem. While LLMs seem to be adept at this sort of step-by-step decomposition, LLMs often make logical and arithmetic mistakes in the solution part, even when the problem is decomposed correctly. In this paper, we present Program-Aided Language models (PAL): a novel approach that uses the LLM to read natural language problems and generate programs as the intermediate reasoning steps, but offloads the solution step to a runtime such as a Python interpreter. With PAL, decomposing the natural language problem into runnable steps remains the only learning task for the LLM, while solving is delegated to the interpreter. We demonstrate this synergy between a neural LLM and a symbolic interpreter across 13 mathematical, symbolic, and algorithmic reasoning tasks from BIG-Bench Hard and others. In all these natural language reasoning tasks, generating code using an LLM and reasoning using a Python interpreter leads to more accurate results than much larger models. For example, PAL using Codex achieves state-of-the-art few-shot accuracy on GSM8K, surpassing PaLM which uses chain-of-thought by absolute 15% top-1. 

Abstract. We present a novel photolytic source of gas-phase NO3 suitable for use in atmospheric chemistry studies that has several advantages over traditional sources that utilize NO2 + O3 reactions and/or thermal dissociation of dinitrogen pentoxide (N2O5). The method generates NO3 via irradiation of aerated aqueous solutions of ceric ammonium nitrate (CAN, (NH4)2Ce(NO3)6) and nitric acid (HNO3) or sodium nitrate (NaNO3). We present experimental and model characterization of the NO3 formation potential of irradiated CAN / HNO3 and CAN / NaNO3 mixtures containing [CAN] = 10−3 to 1.0 M, [HNO3] = 1.0 to 6.0 M, [NaNO3] = 1.0 to 4.8 M, photon fluxes (I) ranging from 6.9 × 1014 to 1.0 × 1016 photons cm−2 s−1, and irradiation wavelengths ranging from 254 to 421 nm. NO3 mixing ratios ranging from parts per billion to parts per million by volume were achieved using this method. At the CAN solubility limit, maximum [NO3] was achieved using [HNO3] ≈ 3.0 to 6.0 M and UVA radiation (λmax⁡ = 369 nm) in CAN / HNO3 mixtures or [NaNO3] ≥ 1.0 M and UVC radiation (λmax⁡ = 254 nm) in CAN / NaNO3 mixtures. Other reactive nitrogen (NO2, N2O4, N2O5, N2O6, HNO2, HNO3, HNO4) and reactive oxygen (HO2, H2O2) species obtained from the irradiation of ceric nitrate mixtures were measured using a NOx analyzer and an iodide-adduct high-resolution time-of-flight chemical ionization mass spectrometer (HR-ToF-CIMS). To assess the applicability of the method for studies of NO3-initiated oxidative aging processes, we generated and measured the chemical composition of oxygenated volatile organic compounds (OVOCs) and secondary organic aerosol (SOA) from the β-pinene + NO3 reaction using a Filter Inlet for Gases and AEROsols (FIGAERO) coupled to the HR-ToF-CIMS.

 

Abstract Energy efficiency improvement is often hindered by the energy efficiency gap. This paper examines the effect of short-run temperature fluctuations on the Energy Star air conditioner purchases in the United States from 2006 to 2019 using transaction-level data. Results show that the probability of purchasing an Energy Star air conditioner increases as the weekly temperature before the transaction deviates from 20–22 °C. A larger response is related to fewer cooling degree days in the previous years, higher electricity prices/income/educational levels/age/rate of owners, more common use of electricity, and stronger concern about climate change. 1 °C increase and decrease from 21 °C would lead to a reduction of total energy expenditure by 35.46 and 17.73 million dollars nationwide (0.13% and 0.06% of the annual total energy expenditure on air conditioning), respectively. Our findings have important policy implications for demand-end interventions to incorporate the potential impact of the ambient physical environment. 

Despite the major advances in NLP, significant disparities in NLP system performance across languages still exist. Arguably, these are due to uneven resource allocation and sub-optimal incentives to work on less resourced languages. To track and further incentivize the global development of equitable language technology, we introduce GlobalBench. Prior multilingual benchmarks are static and have focused on a limited number of tasks and languages. In contrast, GlobalBench is an ever-expanding collection that aims to dynamically track progress on all NLP datasets in all languages. Rather than solely measuring accuracy, GlobalBench also tracks the estimated per-speaker utility and equity of technology across all languages, providing a multi-faceted view of how language technology is serving people of the world. Furthermore, GlobalBench is designed to identify the most under-served languages, and rewards research efforts directed towards those languages. At present, the most under-served languages are the ones with a relatively high population, but nonetheless overlooked by composite multilingual benchmarks (like Punjabi, Portuguese, and Wu Chinese). Currently, GlobalBench covers 966 datasets in 190 languages, and has 1,128 system submissions spanning 62 languages.