Abstract
This paper describes a methodology for constructing and simulating from models of daily weather time series at multiple locations, incorporating potential nonstationarities and suitable for use in those studies of climate impacts and adaptation where a detailed representation of local weather is required. The approach is based on generalised linear models (GLMs) and aims to allow for realistic representations of local weather structures including spatial, temporal and inter-variable dependencies. The theory is implemented in a software tool, Rglimclim, that runs in the R programming environment; and is illustrated using a case study involving generation of daily precipitation and temperature at 26 locations in northern Iberia.
Highlights
The 13th Sustainable Development Goal of the United Nations is ‘take urgent action to combat climate change and its impacts’ (United Nations, 2015)
Apart from methods that seek historical analogs of a given configuration of global climate models (GCMs) outputs, the only downscaling methodologies that meet the requirements of such demanding applications are those based on regional climate models (RCMs) and on weather generators
RCMs are high-resolution climate models that are run for specified regions, usually at a continental scale, and with boundary conditions taken from GCM simulations: modern RCMs can routinely produce output on spatial scales of around 10 × 10 km2, their effective resolution is probably coarser than this (Maraun et al, 2010)
Summary
The 13th Sustainable Development Goal of the United Nations is ‘take urgent action to combat climate change and its impacts’ (United Nations, 2015). A more principled approach is to incorporate indices of large-scale atmospheric structure directly into the weather generator specification In this case the chosen indices must both capture the climate change signal and be well represented by GCMs. For more discussion of this and other considerations in the use of statistical downscaling, see Wilby et al (2004) and Maraun and Widmann (2018, Section 11.5). For applications where a detailed representation of weather inputs is needed, a weather generator must be capable of: producing sequences with realistic spatial, temporal and inter-variable dependence structures over a range of scales; capturing realistic levels of unstructured variability associated with phenomena such as extremes; representing systematic variation over space and time, including under scenarios of climate change; generating sequences at locations for which no data are available; and coping with missing values in the historical records used for model calibration.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
