Robust rescaling methods for integrated water, food, energy security management under uncertainty

Ermolieva, T., Ermoliev, Y., Havlik, P. ORCID: https://orcid.org/0000-0001-5551-5085, Mosnier, A., & Obersteiner, M. ORCID: https://orcid.org/0000-0001-6981-2769 (2013). Robust rescaling methods for integrated water, food, energy security management under uncertainty. Paper presented at [[XIII International Conference on Stochastic Programming]] (ICSP2013), 8-12 July 2013, Bergamo, Italy

Full text not available from this repository.

Abstract

The aim of the paper is to discuss robust non-Bayesian probabilistic cross-entropy based disaggregation (downscaling) techniques driven by the need to address local herterogeneities related to secure food, water, energy provisions consistent with available aggregate data and projections of global and national development trends. For example, aggregate land use projections derived from global economic land use planning models give no insights regarding potentially critical heterogeneities of local processes. High spatial resolution land use and cover change projections are also required as one the crucial inputs into Global Circulation Models. Many practical studies analyzing regional developments use cross-entropy minimization as an underlying principle for estimation of local processes consistent with available aggregate data. Traditional cross-entropy downscaling relies on a single prior distribution. In reality, prior distributions depend on various "environmental" parameters which may not be known exactly. Therefore in general instead of a uniquely defined prior there is a feasible set of these distributions. In this case, the estimation of local changes consistent with available aggregate data can be formulated as probabilistic inverse (from aggregate to local data) problem in the form of, in general, stochastic non-convex cross-entropy minimization model. Specific reparametrization permits to convexify the model. The duality relations derive numerical procedure for local estimates robust with respect to all priors from the feasible set. The approach will be illustrated by downscaling regional GLOBIOM (Global Biosphere Management Mode) model projections of land use changes.

Item Type: Other
Research Programs: Ecosystems Services and Management (ESM)
Bibliographic Reference: Paper presented at [[XIII International Conference on Stochastic Programming]] (ICSP2013), 8-12 July 2013, Bergamo, Italy
Depositing User: IIASA Import
Date Deposited: 15 Jan 2016 08:49
Last Modified: 27 Aug 2021 17:23
URI: https://pure.iiasa.ac.at/10621

Actions (login required)

View Item View Item