Search for: All records

The coastal ocean contributes to regulating atmospheric greenhouse gas concentrations by taking up carbon dioxide (CO₂) and releasing nitrous oxide (N₂O) and methane (CH₄). In this second phase of the Regional Carbon Cycle Assessment and Processes (RECCAP2), we quantify global coastal ocean fluxes of CO₂, N₂O and CH₄using an ensemble of global gap‐filled observation‐based products and ocean biogeochemical models. The global coastal ocean is a net sink of CO₂in both observational products and models, but the magnitude of the median net global coastal uptake is ∼60% larger in models (−0.72 vs. −0.44 PgC year⁻¹, 1998–2018, coastal ocean extending to 300 km offshore or 1,000 m isobath with area of 77 million km²). We attribute most of this model‐product difference to the seasonality in sea surface CO₂partial pressure at mid‐ and high‐latitudes, where models simulate stronger winter CO₂uptake. The coastal ocean CO₂sink has increased in the past decades but the available time‐resolving observation‐based products and models show large discrepancies in the magnitude of this increase. The global coastal ocean is a major source of N₂O (+0.70 PgCO₂‐e year⁻¹in observational product and +0.54 PgCO₂‐e year⁻¹in model median) and CH₄(+0.21 PgCO₂‐e year⁻¹in observational product), which offsets a substantial proportion of the coastal CO₂uptake in the net radiative balance (30%–60% in CO₂‐equivalents), highlighting the importance of considering the three greenhouse gases when examining the influence of the coastal ocean on climate.

 

In this work, we revisit the classical stochastic jump-diffusion process and develop an effective variant for estimating visibility statuses of objects while tracking them in videos. Dealing with partial or full occlusions is a long standing problem in computer vision but largely remains unsolved. In this work, we cast the above problem as a Markov Decision Process and develop a policy-based jump-diffusion method to jointly track object locations in videos and estimate their visibility statuses. Our method employs a set of jump dynamics to change object’s visibility statuses and a set of diffusion dynamics to track objects in videos. Different from traditional jump-diffusion process that stochastically generates dynamics, we utilize deep policy functions to determine the best dynamic for the present state and learn the optimal policies using reinforcement learning methods. Our method is capable of tracking objects with full or partial occlusions in crowded scenes. We evaluate the proposed method over challenging video sequences and compare it to alternative tracking methods. Significant improvements are made particularly for videos with frequent interactions or occlusions. 

Abstract In the quest for high-energy neutrino sources, the Astrophysical Multimessenger Observatory Network has implemented a new search by combining data from the High Altitude Water Cherenkov (HAWC) Observatory and the Astronomy with a Neutrino Telescope and Abyss environmental RESearch (ANTARES) neutrino telescope. Using the same analysis strategy as in a previous detector combination of HAWC and IceCube data, we perform a search for coincidences in HAWC and ANTARES events that are below the threshold for sending public alerts in each individual detector. Data were collected between 2015 July and 2020 February with a live time of 4.39 yr. Over this time period, three coincident events with an estimated false-alarm rate of <1 coincidence per year were found. This number is consistent with background expectations. 

Abstract For several decades, the origin of ultra-high-energy cosmic rays (UHECRs) has been an unsolved question of high-energy astrophysics. One approach for solving this puzzle is to correlate UHECRs with high-energy neutrinos, since neutrinos are a direct probe of hadronic interactions of cosmic rays and are not deflected by magnetic fields. In this paper, we present three different approaches for correlating the arrival directions of neutrinos with the arrival directions of UHECRs. The neutrino data are provided by the IceCube Neutrino Observatory and ANTARES, while the UHECR data with energies above ∼50 EeV are provided by the Pierre Auger Observatory and the Telescope Array. All experiments provide increased statistics and improved reconstructions with respect to our previous results reported in 2015. The first analysis uses a high-statistics neutrino sample optimized for point-source searches to search for excesses of neutrino clustering in the vicinity of UHECR directions. The second analysis searches for an excess of UHECRs in the direction of the highest-energy neutrinos. The third analysis searches for an excess of pairs of UHECRs and highest-energy neutrinos on different angular scales. None of the analyses have found a significant excess, and previously reported overfluctuations are reduced in significance. Based on these results, we further constrain the neutrino flux spatially correlated with UHECRs.