Post-processing Climate Projection Output for Key Users in Norway

For long-term planning, possible future climate changes, as well as their consequences, should be considered, whenever relevant for decision-making. The relevance of climate information, however, varies from sector to sector. In PostClim, the Norwegian Centre for Climate Services (NCCS) and the Norwegian Computing Center (NR) are assisted by decision-makers from municipalities and from the agricultural sector in order to define what kind of climate information is important for them. The main purpose of PostClim is to find how we best can produce and distribute information on climate change and hydrological consequences that is tailored to the needs in these sectors. The municipalities are represented by the municipality sector's organization KS, and the agricultural sector by the Norwegian Farmers' Union, the Federation of Norwegian Agricultural Cooperatives, Norwegian Agricultural Advisory Board and the Norwegian Institute of Bioeconomy Research (NIBIO). Local climate projections are based on regional or global climate models. Data from such models provide useful climate information on a larger scale, but usually, they cannot be used directly in a given locality, because of model errors and because the real landscape is more complex than the landscape in the models. The model data are therefore post-processed by various methods with different strengths and weaknesses. A key issue in PostClim has been to validate and further develop such methods, with particular regard to the needs of the decision-makers. A flexible evaluation tool has been developed to provide ranking of climate model data. The tool is programmed in R, and freely available. Feedback from the PostClim key users included priorities, both concerning variables and critical qualities for the post-processing methods. Temperature and precipitation were listed as the two most important meteorological variables to post-process, but also air humidity, evaporation, wind speed and global radiation are important, e.g. for modelling hydrological impacts of climate change. At least daily time resolution is requested and a spatial resolution of 1x1 km. So far, only projections of daily temperature and precipitation have been available for Norway at this fine resolution. In PostClim fine-scale projections for daily max- and min-temperature, and for daily wind average were produced based upon the same models (EURO-CORDEX) that were used in the report ?Klima i Norge 2100? from 2015. Data for max- and min-temperature are available from klimaservicesenter.no along with other 1x1 km projection data. These data were used to produce maps showing projections for frost-change-days. The municipalities as well as the agriculture sector have requested such projections. Fine scale wind projections are available for research at https://thredds.met.no/ (contact KSS or MET). Feedback from the user panel further confirmed that changes in drought and flooding conditions (including stream floods and urban floods) are important both for municipalities and the agriculture sector. PostClim therefore includes hydrological modeling, and has a particular focus on establishing post-processing techniques that produce realistic input data on the temporal and spatial scales needed for such models. In PostClim, two methods are applied to establish temperature and precipitation datasets at 1x1 km spatial scale that preserves intra-variable consistency in space and time, as well as inter-variable consistency. One is presently under development in a PhD-project, which is scheduled to conclude in the summer of 2020. This method is very promising, but also resource demanding, so it is being tested only in Trøndelag county. The other method consists of redistribution of the daily values from existing projections. The shuffled data will be available either at https://thredds.met.no/ or at the NCCS website. For further evaluation of post-processing methods and user defined climate change impact products, a spatially distributed hydrological model with radiation based evaporation and snowmelt scheme has been developed and set up for Norway with 1x1 km2 spatial resolution. The model simulations represent an improvement in simulating land-surface-atmosphere interactions and land-surface hydrological processes including streamflow. The Norwegian Environmental Agency has signaled that they will commission a new national climate report for Norway in the wake of the new main report from IPCC. Several tools developed in PostClim will then be valuable. The shuffling method for improving the consistency in the projection data, the improved descriptions of evaporation and snow in the hydrological model, and the validation tool for post-processing techniques will probably all be useful.

Det er i prosjektet publisert finskala klimaframskrivninger for døgnlig maksimum og minimumtemperatur, antall dager med nullgradspasseringer og for vind. Dette er viktige variabel både for planlegning av infrastruktur, og for vurdering av klimaendringenes konsekvenser for landbruk og for økologiske systemer. Det forventes at disse vil bli brukt i modeller for virkninger av klimaendringer. Det er også utviklet og implementert nye metoder som planlegges brukt ved produksjonen av neste klimarapport for Norge, som forventes å komme i kjølvannet av neste hovedrapport fra FNs klimapanel. Dette gjelder både metoder for å bevare konsistens i datasettene for temperatur og nedbør, og ny formuleringen av fordampning i den hydrologiske modellen. Prosjektet bidrar på denne måten til bedre kunnskap om konsekvenser av klimaendringer i Norge, og derfor bedre forutsetninger for klimatilpasning.

Long-term planning and decision-making in the context of climate impacts, adaptation and vulnerability requires information on future climate and potential risks that are tailored to the individual stakeholder's needs. In PostClim, partners from the Norwegian Centre for Climate Services will join forces with decision-makers from Norwegian municipalities and the agricultural sector to produce appropriate climate projections for decision-making under an uncertain future climate. To this aim, post-processing techniques for statistical downscaling and bias adjustment will be employed, and further developed as necessary, to bridge the gap between the raw output from global and regional climate model and the needs of the decision-makers. A number of different techniques will be validated according to criteria that are important for the stakeholders. The further development of the methods will also be based upon user needs. Earlier experiences indicated that the impact of climate change on stream floods and the flooding pattern in general are import and for the planning of infrastructure. PostClim thus includes hydrological modelling.