BZPEERReading, trust and collaboration for researchDiscover
Workspaces
Network
Opportunities
Account
Researcher profile
Niklas Brännström contributes to research discovery and scholarly infrastructure.
Trust snapshot
Actions
Research graph
Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.
BZPEER is loading the nearby papers, people, topics and institutions for this page.
Published work
We show that the lower and upper Frechét-Hoeffding copulas, which are singular, can be regularized to absolutely continuous copulas. The method, which is constructive and explicit, states sufficient conditions for when an absolutely continuous copula can be achieved by averaging. A higher degree of regularisation cannot be achieved with the proposed method.
We consider a bilevel optimatisation method for inverse linear atmospheric dispersion problems where both linear and non-linear model parameters are to be determined. We propose that a smooth weighted Mahalanobis distance function is used and derive sufficient conditions for when the follower problem has local strict convexity. A few toy-models are presented where local strict convexity and ill-posedness of the inverse problem are explored, indeed the smooth distance function is compared and contrasted to linear and piecewise linear ones. The bilevel optimisation method is then applied to sensor data collected in wind tunnel experiments of a neutral gas release in urban environments (MODITIC).
On September 9 2014 several incidences of foul smell (rotten eggs) were reported on the coast of Norway (in particular in the vicinity of Molde) and then on September 10 in the interior parts of county Västerbotten, Sweden. One of the theories that were put forward was that the foul smell was due to degassing of the Bardarbunga volcano on Iceland. Using satellite images (GOME-1,-2) of the sulphur dioxide, SO_2, contents in the atmosphere surrounding Iceland to estimate flux of SO_2 from the volcano and an atmospheric transport model, PELLO, we vindicate this theory: we argue that the cause for the foul smell was hydrogen sulphide originating from Bardarbunga. The model concentrations are also compared to SO_2 concentration measurements from Muonio, Finland.
Using measure theoretic arguments, we provide a general framework for describing and studying the general linear inverse dispersion problem where no a priori assumptions on the source function has been made. We investigate the source-sensor relationship and rigorously state solvability conditions for when the inverse problem can be solved using a least-squares optimisation method. That is, we derive conditions for when the least-squares problem is well-defined.
Footprint analysis, also known as the study of Influence areas, is a first order method for solving inverse atmospheric dispersion problems. We revisit the concept of footprints giving a rigorous definition of the concept (denoted posterior footprints and posterior zero footprints) in terms of spatio-temporal domains. The notion of footprints is then augmented the to the forward dispersion problem by defining prior footprints and prior zero footprints. We then study how posterior footprints and posterior zero footprints can be combined to reveal more information about the source, and how prior footprints and prior footprints can be combined to yield more information about the measurements.
We study a class of slow-fast Hamiltonian systems with any finite number of degrees of freedom, but with at least one slow one and two fast ones. At $% ε=0$ the slow dynamics is frozen. We assume that the frozen system (i.e. the unperturbed fast dynamics) has families of hyperbolic periodic orbits with transversal heteroclinics. For each periodic orbit we define an action $J.$ This action may be viewed as an action Hamiltonian (in the slow variables). It has been shown in (N. Brännström, V.Gelfreich (2008)) that there are orbits of the full dynamics which shadow any \emph{finite} combination of forward orbits of $J$ for a time $t=O(ε^{-1})$. We introduce an assumption on the mutual relationship between the actions $J.$ This assumption enables us to shadow any continuous curve (of arbitrary length) in the slow phase space for any time. The slow dynamics shadows the curve as a purely geometrical object, thus the time on the slow dynamics has to be reparameterised.