Performance of the Adriatic Sea and Coast (AdriSC) climate component – a COAWST V3.3-based coupled atmosphere–ocean modelling suite: atmospheric dataset Performance of the Adriatic Sea and Coast (AdriSC) climate component – a COAWST V3.3-based... Cléa Denamiel et al.
In this evaluation study, the coupled atmosphere–ocean Adriatic Sea and Coast (AdriSC) climate model, which was implemented to carry out 31-year evaluation and climate projection simulations in the Adriatic and northern Ionian seas, is briefly presented. The kilometre-scale AdriSC atmospheric results, derived with the Weather Research and Forecasting (WRF) 3 km model for the 1987–2017 period, are then thoroughly compared to a comprehensive publicly and freely available observational dataset. The evaluation shows that overall, except for the summer surface temperatures, which are systematically underestimated, the AdriSC WRF 3 km model has a far better capacity to reproduce surface climate variables (and particularly the rain) than the WRF regional climate models at 0.11∘ resolution. In addition, several spurious data have been found in both gridded products and in situ measurements, which thus should be used with care in the Adriatic region for climate studies at local and regional scales. Long-term simulations with the AdriSC climate model, which couples the WRF 3 km model with a 1 km ocean model, might thus be a new avenue to substantially improve the reproduction, at the climate scale, of the Adriatic Sea dynamics driving the Eastern Mediterranean thermohaline circulation. As such it may also provide new standards for climate studies of orographically developed coastal regions in general.
Denamiel, C., Pranić, P., Ivanković, D., Tojčić, I., and Vilibić, I.: Performance of the Adriatic Sea and Coast (AdriSC) climate component – a COAWST V3.3-based coupled atmosphere–ocean modelling suite: atmospheric dataset, Geosci. Model Dev., 14, 3995–4017, https://doi.org/10.5194/-14-3995-2021, 2021.
Zagreb, Croatia 14 Day Weather Forecast
In the past decade, within the climate community scientific efforts led by the COordinated Regional climate Downscaling EXperiment (CORDEX; Giorgi et al., 2009) facilitated the rapid development and applications of Regional Climate Models (RCMs) around the world (e.g. Rinke etal. 2011; Nikulin et al., 2012; da Rocha et al., 2014; Huang et al., 2015; Ruti et al., 2016; Zou and Zhou, 2017; Di Virgilio et al., 2019). Specifically, in the Mediterranean Sea, several RCMs have been developed within the Med-CORDEX initiative (e.g. Sevault et al., 2014; Ruti et al., 2016; Somot et al., 2018; Reale et al., 2020; Sein et al., 2020). However, RCMs often fail to represent extreme events as, for example, they do not properly resolve complex orography, coastline and land–sea contrasts (Prein et al., 2015). Consequently, the need to study climatic hazards and their extremes with kilometre-scale atmospheric models has recently been promoted (e.g. the summer 2020 call for the CORDEX Flagship Pilot Study, https://cordex.org/wp-content/uploads/2020/07/FPS-flyer-summer2020.pdf, last access: 26 June 2021). Additionally, in coastal regions, such atmospheric models should be coupled with high-resolution ocean models in order to quantify the impact of these extreme conditions on the ocean dynamics and therefore on the marine ecosystems, the erosion or the transport of pollutants, or other systems. However, due to their prohibitive numerical cost, coupled atmosphere–ocean kilometre-scale climate models are not commonly used in long-term climate research.
Figure 1Name of the geographical (a) and orographic and bathymetric (b) features of the AdriSC WRF 3 km model domain, location of the UWYO soundings (b), and biases between the AdriSC WRF 3 km orography and both the NOAA stations (c) and the E-OBS dataset (d) elevations.
Nevertheless, over the elongated semi-enclosed Adriatic basin (Fig. 1), only high-resolution limited-area models can represent the atmosphere–ocean interactions during extreme events (e.g. Pasarić et al., 2007; Prtenjak et al., 2010; Ricchi et al., 2016; Cavaleri et al., 2010, 2018). The complex geomorphology of the Adriatic Sea indeed includes (a) more than 1200 islands, islets, ridges, and rocks, mostly located along the northeastern coastline; (b) mountains surrounding the entire basin; and (c) bathymetries evolving from a really shallow and wide shelf in the north to a deep pit in the south. Additionally, orographically driven extreme windstorms mostly from the northeastern direction (i.e. the so-called bora winds; Brzović and Strelec Mahović, 1999; Grisogono and Belušić 2009) are known to strongly influence the annual dense water budget in the Adriatic Sea. The dense waters are formed on both northern Adriatic shelf (through shallow-water cooling, Janeković et al., 2014) and in the deep southern Adriatic (through open-ocean convection, Gačić et al., 2002) and are a driver of interannual to decadal thermohaline and biogeochemical variability between the Adriatic and the northern Ionian seas (Roether and Schlitzer, 1991; Gačić et al., 2010; Bensi et al., 2013; Batistić et al., 2014). The Adriatic Sea and Coast (AdriSC) kilometre-scale climate model (Denamiel et al., 2019) was thus recently developed to represent the long-term atmospheric and oceanic circulations in the Adriatic basin, with the perspective of future applications to extreme event hazard assessment, ecosystem modelling, sediment and larvae transport, etc. Furthermore, for climate projections, the pseudo-global warming (PGW) downscaling method (Schär et al., 1996) was proven to greatly improve the future projections of precipitation and convective storms in atmospheric kilometre-scale climate models (Prein et al., 2015). Consequently, this method was first extended to coupled atmosphere–ocean models, then implemented in the AdriSC climate component and finally tested successfully with an ensemble of short-term climate simulations for wind-driven extreme events in the Adriatic Sea (Denamiel et al., 2020a, b). The need to use kilometre-scale models during severe bora events for proper representation of the Adriatic long-term thermohaline circulation was also demonstrated (Denamiel et al., 2021).
Zagreb Croatia, 14 Days Weather Forecast, Radar & Photos
Following these preliminary results and the high-resolution studies implemented in other parts of the world (e.g. Chan et al., 2018; Li et al., 2019; Knist et al., 2020), a 31-year evaluation run was performed with the AdriSC climate model for the 1987–2017 period. It should be noted that, in 2018 when the climate model was set up, the 1987–2017 period was chosen due to the availability of reliable daily ocean re-analysis in the Mediterranean Sea. Additionally, contrary to global or regional climate models, the evaluation of kilometre-scale models requires the use of observational datasets with high temporal resolution (i.e. at least hourly in the atmosphere and daily in the ocean) and spatial coverage (i.e. a network of in situ measurements or kilometre-scale gridded products) for both atmospheric and oceanic essential climate variables. However, such datasets are intrinsically uncertain and therefore not entirely reliable. For example, (a) ground-based station measurements often present inhomogeneities due to change in instruments or environmental conditions; (b) long-term time series are difficult to obtain from measurements at sea that highly depend on the ship and buoy locations; and (c) remote sensing data generally have low temporal resolution (i.e. daily), do not measure the climate variables directly, and can include systemic disturbances due to the impact of the atmosphere. Moreover, based on the assumption that the quality of the observational datasets can be assessed with climate models, Massonnet et al. (2016) highlighted the need to provide guidance for a more objective selection of the observations used in evaluation studies. Another key point concerning the atmospheric observational datasets, and most particularly the in situ measurements, is the ease (and cost) of access, which highly depends on the data sharing policy of the different providers. For example, collecting ground-based station data from the different meteorological agencies around the Adriatic basin requires contacting each provider separately (Italy, Croatia, Montenegro, Albania, etc.), and in many cases this implies receiving, after a long delay, partial datasets provided in different formats. Consequently, only open-access datasets accessible on the web and provided in an easily readable format are used hereafter for the evaluation of the AdriSC climate component. The inconvenience of this choice is that the quality of the datasets can sometimes be degraded before being shared online due to, for example, unit conversions and rounding errors.
The following study solely assesses the skill of the AdriSC atmospheric kilometre-scale model, while the evaluation of the AdriSC ocean coastal model is done separately. It is also , as suggested by Massonnet et al. (2016), a bidirectional exercise evaluating both the kilometre-scale AdriSC atmospheric model and the freely available observations retrieved in the Adriatic basin from in situ measurements, gridded datasets and remote-sensing products. The presented work thus aims at answering the following questions: what are the strengths and shortcomings of the AdriSC atmospheric model depending on the evaluated essential climate variables, and how are they related to the physical set-up of the model? Are the skills of the newly developed climate model similar at daily and hourly timescales? How does the performance of the kilometre-scale atmospheric model compare to the RCMs set up within the CORDEX community? What is the quality and the reliability of the freely available observations in the Adriatic region?
Consequently, the AdriSC modelling suite (i.e. its climate component, its web portal and the set-up of its atmospheric model) and the observations and methods used to perform the skill assessment of the model are first presented in Sect. 2. Following this, Sect. 3 displays and discusses the results, consisting of basic, spatially distributed statistical or seasonal and vertical skill assessments, as well as comparisons of measured and modelled climatologies and distributions. Finally, the findings of this study are summarised in Sect. 4.
Phenolic Compounds In Rosaceae Fruit And Nut Crops
The Adriatic Sea and Coast (AdriSC) modelling suite (Denamiel et al. 2019) has been developed with the aim of accurately representing the processes driving atmospheric and oceanic circulation at different temporal and spatial scales over the Adriatic and northern
The Adriatic Sea and Coast (AdriSC) modelling suite (Denamiel et al. 2019) has been developed with the aim of accurately representing the processes driving atmospheric and oceanic circulation at different temporal and spatial scales over the Adriatic and northern
