Human impacts and their interactions in the Baltic Sea region

Coastal environments, in particular heavily populated semi-enclosed marginal seas and coasts like the Baltic Sea region, are strongly affected by human activities. A multitude of human impacts, including climate change, affect the different compartments of the environment, and these effects interact with each other. As part of the Baltic Earth Assessment Reports (BEAR), we present an inventory and discussion of different human-induced factors and processes affecting the environment of the Baltic Sea region, and their interrelations. Some are naturally occurring and modified by human activities (i.e. climate change, coastal processes, hypoxia, acidification, submarine groundwater discharges, marine ecosystems, non-indigenous species, land use and land cover), some are completely human-induced (i.e. agriculture, aquaculture, fisheries, river regulations, offshore wind farms, shipping, chemical contamination, dumped warfare agents, marine litter and microplastics, tourism, and coastal management), and they are all interrelated to different degrees. We present a general description and analysis of the state of knowledge on these interrelations. Our main insight is that climate change has an overarching, integrating impact on all of the other factors and can be interpreted as a background effect, which has different implications for the other factors. Impacts on the environment and the human sphere can be roughly allocated to anthropogenic drivers such as food production, energy production, transport, industry and economy. The findings from this inventory of available information and analysis of the different factors and their interactions in the Baltic Sea region can largely be transferred to other comparable marginal and coastal seas in the world.

Balanced estimate and uncertainty assessment of European climate change using the large EURO-CORDEX regional climate model ensemble

Large multiscenario multimodel ensembles (MMEs) of regional climate model (RCM) experiments driven by global climate models (GCMs) are made available worldwide and aim at providing robust estimates of climate changes and associated uncertainties. Due to many missing combinations of emission scenarios and climate models leading to sparse scenario–GCM–RCM matrices, these large ensembles, however, are very unbalanced, which makes uncertainty analyses impossible with standard approaches. In this paper, the uncertainty assessment is carried out by applying an advanced statistical approach, called QUALYPSO, to a very large ensemble of 87 EURO-CORDEX climate projections, the largest MME based on regional climate models ever produced in Europe. This analysis provides a detailed description of this MME, including (i) balanced estimates of mean changes for near-surface temperature and precipitation in Europe, (ii) the total uncertainty of projections and its partition as a function of time, and (iii) the list of the most important contributors to the model uncertainty. For changes in total precipitation and mean temperature in winter (DJF) and summer (JJA), the uncertainty due to RCMs can be as large as the uncertainty due to GCMs at the end of the century (2071–2099). Both uncertainty sources are mainly due to a small number of individual models clearly identified. Due to the highly unbalanced character of the MME, mean estimated changes can drastically differ from standard average estimates based on the raw ensemble of opportunity. For the RCP4.5 emission scenario in central–eastern Europe for instance, the difference between balanced and direct estimates is up to 0.8 ∘C for summer temperature changes and up to 20 % for summer precipitation changes at the end of the century.

Ubiquity of human-induced changes in climate variability

While climate change mitigation targets necessarily concern maximum mean state changes, understanding impacts and developing adaptation strategies will be largely contingent on how climate variability responds to increasing anthropogenic perturbations. Thus far Earth system modeling efforts have primarily focused on projected mean state changes and the sensitivity of specific modes of climate variability, such as the El Niño–Southern Oscillation. However, our knowledge of forced changes in the overall spectrum of climate variability and higher-order statistics is relatively limited. Here we present a new 100-member large ensemble of climate change projections conducted with the Community Earth System Model version 2 over 1850–2100 to examine the sensitivity of internal climate fluctuations to greenhouse warming. Our unprecedented simulations reveal that changes in variability, considered broadly in terms of probability distribution, amplitude, frequency, phasing, and patterns, are ubiquitous and span a wide range of physical and ecosystem variables across many spatial and temporal scales. Greenhouse warming in the model alters variance spectra of Earth system variables that are characterized by non-Gaussian probability distributions, such as rainfall, primary production, or fire occurrence. Our modeling results have important implications for climate adaptation efforts, resource management, seasonal predictions, and assessing potential stressors for terrestrial and marine ecosystems.

How can solar geoengineering and mitigation be combined under climate targets?

So far, scientific analyses have mainly focused on the pros and cons of solar geoengineering or solar radiation management (SRM) as a climate policy option in mere isolation. Here, we put SRM into the context of mitigation by a strictly temperature-target-based approach. As the main innovation, we present a scheme that extends the applicability regime of temperature targets from mitigation-only to SRM-mitigation analyses. We explicitly account for one major category of side effects of SRM while minimizing economic costs for complying with the 2 ∘C temperature target. To do so, we suggest regional precipitation guardrails that are compatible with the 2 ∘C target. Our analysis shows that the value system enshrined in the 2 ∘C target leads to an elimination of most of the SRM from the policy scenario if a transgression of environmental targets is confined to 1/10 of the standard deviation of natural variability. Correspondingly, about half to nearly two-thirds of mitigation costs could be saved, depending on the relaxation of the precipitation criterion. In addition, assuming a climate sensitivity of 3 ∘C or more, in case of a delayed enough policy, a modest admixture of SRM to the policy portfolio might provide debatable trade-offs compared to a mitigation-only future. Also, in our analysis which abstains from a utilization of negative emissions technologies, for climate sensitivities higher than 4 ∘C, SRM will be an unavoidable policy tool to comply with the temperature targets. The economic numbers we present must be interpreted as upper bounds in the sense that cost-lowering effects by including negative emissions technologies are absent. However, with an additional climate policy option such as carbon dioxide removal present, the role of SRM would be even more limited. Hence, our results, pointing to a limited role of SRM in a situation of immediate implementation of a climate policy, are robust in that regard. This limitation would be enhanced if further side effects of SRM are taken into account in a target-based integrated assessment of SRM.

Extreme metrics from large ensembles: investigating the effects of ensemble size on their estimates

We consider the problem of estimating the ensemble sizes required to characterize the forced component and the internal variability of a number of extreme metrics. While we exploit existing large ensembles, our perspective is that of a modeling center wanting to estimate a priori such sizes on the basis of an existing small ensemble (we assume the availability of only five members here). We therefore ask if such a small-size ensemble is sufficient to estimate accurately the population variance (i.e., the ensemble internal variability) and then apply a well-established formula that quantifies the expected error in the estimation of the population mean (i.e., the forced component) as a function of the sample size n, here taken to mean the ensemble size. We find that indeed we can anticipate errors in the estimation of the forced component for temperature and precipitation extremes as a function of n by plugging into the formula an estimate of the population variance derived on the basis of five members. For a range of spatial and temporal scales, forcing levels (we use simulations under Representative Concentration Pathway 8.5) and two models considered here as our proof of concept, it appears that an ensemble size of 20 or 25 members can provide estimates of the forced component for the extreme metrics considered that remain within small absolute and percentage errors. Additional members beyond 20 or 25 add only marginal precision to the estimate, and this remains true when statistical inference through extreme value analysis is used. We then ask about the ensemble size required to estimate the ensemble variance (a measure of internal variability) along the length of the simulation and – importantly – about the ensemble size required to detect significant changes in such variance along the simulation with increased external forcings. Using the F test, we find that estimates on the basis of only 5 or 10 ensemble members accurately represent the full ensemble variance even when the analysis is conducted at the grid-point scale. The detection of changes in the variance when comparing different times along the simulation, especially for the precipitation-based metrics, requires larger sizes but not larger than 15 or 20 members. While we recognize that there will always exist applications and metric definitions requiring larger statistical power and therefore ensemble sizes, our results suggest that for a wide range of analysis targets and scales an effective estimate of both forced component and internal variability can be achieved with sizes below 30 members. This invites consideration of the possibility of exploring additional sources of uncertainty, such as physics parameter settings, when designing ensemble simulations.

Storylines of weather-induced crop failure events under climate change

Unfavourable weather is a common cause for crop failures all over the world. Whilst extreme weather conditions may cause extreme impacts, crop failure commonly is induced by the occurrence of multiple and combined anomalous meteorological drivers. For these cases, the explanation of conditions leading to crop failure is complex, as the links connecting weather and crop yield can be multiple and non-linear. Furthermore, climate change is likely to perturb the meteorological conditions, possibly altering the occurrences of crop failures or leading to unprecedented drivers of extreme impacts. The goal of this study is to identify important meteorological drivers that cause crop failures and to explore changes in crop failures due to global warming. For that, we focus on a historical failure event, the extreme low soybean production during the 2012 season in the midwestern US. We first train a random forest model to identify the most relevant meteorological drivers of historical crop failures and to predict crop failure probabilities. Second, we explore the influence of global warming on crop failures and on the structure of compound drivers. We use large ensembles from the EC-Earth global climate model, corresponding to present-day, pre-industrial +2 and 3 ∘C warming, respectively, to isolate the global warming component. Finally, we explore the meteorological conditions inductive for the 2012 crop failure and construct analogues of these failure conditions in future climate settings. We find that crop failures in the midwestern US are linked to low precipitation levels, and high temperature and diurnal temperature range (DTR) levels during July and August. Results suggest soybean failures are likely to increase with climate change. With more frequent warm years due to global warming, the joint hot–dry conditions leading to crop failures become mostly dependent on precipitation levels, reducing the importance of the relative compound contribution. While event analogues of the 2012 season are rare and not expected to increase, impact analogues show a significant increase in occurrence frequency under global warming, but for different combinations of the meteorological drivers than experienced in 2012. This has implications for assessment of the drivers of extreme impact events.

Process-based analysis of terrestrial carbon flux predictability

Despite efforts to decrease the discrepancy between simulated and observed terrestrial carbon fluxes, the uncertainty in trends and patterns of the land carbon fluxes remains high. This difficulty raises the question of the extent to which the terrestrial carbon cycle is predictable and which processes explain the predictability. Here, the perfect model approach is used to assess the potential predictability of net primary production (NPPpred) and heterotrophic respiration (Rhpred) by using ensemble simulations conducted with the Max Planck Institute Earth system model. In order to assess the role of local carbon flux predictability (CFpred) in the predictability of the global carbon cycle, we suggest a new predictability metric weighted by the amplitude of the flux anomalies. Regression analysis is used to determine the contribution of the predictability of different environmental drivers to NPPpred and Rhpred (soil moisture, air temperature, and radiation for NPP, and soil organic carbon, air temperature, and precipitation for Rh). Global NPPpred is driven to 62 % and 30 % by the predictability of soil moisture and temperature, respectively. Global Rhpred is driven to 52 % and 27 % by the predictability of soil organic carbon and temperature, respectively. The decomposition of predictability shows that the relatively high Rhpred compared to NPPpred is due to the generally high predictability of soil organic carbon. The seasonality in NPPpred and Rhpred patterns can be explained by the change in limiting factors over the wet and dry months. Consequently, CFpred is controlled by the predictability of the currently limiting environmental factor. Differences in CFpred between ensemble simulations can be attributed to the occurrence of wet and dry years, which influences the predictability of soil moisture and temperature. This variability of predictability is caused by the state dependency of ecosystem processes. Our results reveal the crucial regions and ecosystem processes to be considered when initializing a carbon prediction system.

Impacts of compound hot–dry extremes on US soybean yields

The US agriculture system supplies more than one-third of globally traded soybean, and with 90 % of US soybean produced under rainfed agriculture, soybean trade is particularly sensitive to weather and climate variability. Average growing season climate conditions can explain about one-third of US soybean yield variability. Additionally, crops can be sensitive to specific short-term weather extremes, occurring in isolation or compounding at key moments throughout crop development. Here, we identify the dominant within-season climate drivers that can explain soybean yield variability in the US, and we explore the synergistic effects between drivers that can lead to severe impacts. The study combines weather data from reanalysis and satellite-informed root zone soil moisture fields with subnational crop yields using statistical methods that account for interaction effects. On average, our models can explain about two-thirds of the year-to-year yield variability (70 % for all years and 60 % for out-of-sample predictions). The largest negative influence on soybean yields is driven by high temperature and low soil moisture during the summer crop reproductive period. Moreover, due to synergistic effects, heat is considerably more damaging to soybean crops during dry conditions and is less problematic during wet conditions. Compounding and interacting hot and dry (hot–dry) summer conditions (defined by the 95th and 5th percentiles of temperature and soil moisture respectively) reduce yields by 2 standard deviations. This sensitivity is 4 and 3 times larger than the sensitivity to hot or dry conditions alone respectively. Other relevant drivers of negative yield responses are lower temperatures early and late in the season, excessive precipitation in the early season, and dry conditions in the late season. We note that the sensitivity to the identified drivers varies across the spatial domain. Higher latitudes, and thus colder regions, are positively affected by high temperatures during the summer period. On the other hand, warmer southeastern regions are positively affected by low temperatures during the late season. Historic trends in identified drivers indicate that US soybean production has generally benefited from recent shifts in weather except for increasing rainfall in the early season. Overall, warming conditions have reduced the risk of frost in the early and late seasons and have potentially allowed for earlier sowing dates. More importantly, summers have been getting cooler and wetter over the eastern US. Nevertheless, despite these positive changes, we show that the frequency of compound hot–dry summer events has remained unchanged over the 1946–2016 period. In the longer term, climate models project substantially warmer summers for the continental US, although uncertainty remains as to whether this will be accompanied by drier conditions. This highlights a critical element to explore in future studies focused on US agricultural production risk under climate change.

Exploring the coupled ocean and atmosphere system with a data science approach applied to observations from the Antarctic Circumnavigation Expedition

The Southern Ocean is a critical component of Earth's climate system, but its remoteness makes it challenging to develop a holistic understanding of its processes from the small scale to the large scale. As a result, our knowledge of this vast region remains largely incomplete. The Antarctic Circumnavigation Expedition (ACE, austral summer 2016/2017) surveyed a large number of variables describing the state of the ocean and the atmosphere, the freshwater cycle, atmospheric chemistry, and ocean biogeochemistry and microbiology. This circumpolar cruise included visits to 12 remote islands, the marginal ice zone, and the Antarctic coast. Here, we use 111 of the observed variables to study the latitudinal gradients, seasonality, shorter-term variations, geographic setting of environmental processes, and interactions between them over the duration of 90 d. To reduce the dimensionality and complexity of the dataset and make the relations between variables interpretable we applied an unsupervised machine learning method, the sparse principal component analysis (sPCA), which describes environmental processes through 14 latent variables. To derive a robust statistical perspective on these processes and to estimate the uncertainty in the sPCA decomposition, we have developed a bootstrap approach. Our results provide a proof of concept that sPCA with uncertainty analysis is able to identify temporal patterns from diurnal to seasonal cycles, as well as geographical gradients and “hotspots” of interaction between environmental compartments. While confirming many well known processes, our analysis provides novel insights into the Southern Ocean water cycle (freshwater fluxes), trace gases (interplay between seasonality, sources, and sinks), and microbial communities (nutrient limitation and island mass effects at the largest scale ever reported). More specifically, we identify the important role of the oceanic circulations, frontal zones, and islands in shaping the nutrient availability that controls biological community composition and productivity; the fact that sea ice controls sea water salinity, dampens the wave field, and is associated with increased phytoplankton growth and net community productivity possibly due to iron fertilisation and reduced light limitation; and the clear regional patterns of aerosol characteristics that have emerged, stressing the role of the sea state, atmospheric chemical processing, and source processes near hotspots for the availability of cloud condensation nuclei and hence cloud formation. A set of key variables and their combinations, such as the difference between the air and sea surface temperature, atmospheric pressure, sea surface height, geostrophic currents, upper-ocean layer light intensity, surface wind speed and relative humidity played an important role in our analysis, highlighting the necessity for Earth system models to represent them adequately. In conclusion, our study highlights the use of sPCA to identify key ocean–atmosphere interactions across physical, chemical, and biological processes and their associated spatio-temporal scales. It thereby fills an important gap between simple correlation analyses and complex Earth system models. The sPCA processing code is available as open-access from the following link: https://renkulab.io/gitlab/ACE-ASAID/spca-decomposition (last access: 29 March 2021). As we show here, it can be used for an exploration of environmental data that is less prone to cognitive biases (and confirmation biases in particular) compared to traditional regression analysis that might be affected by the underlying research question.

Reduced-complexity model for the impact of anthropogenic CO₂ emissions on future glacial cycles

We propose a reduced-complexity process-based model for the long-term evolution of the global ice volume, atmospheric CO2 concentration, and global mean temperature. The model's only external forcings are the orbital forcing and anthropogenic CO2 cumulative emissions. The model consists of a system of three coupled non-linear differential equations representing physical mechanisms relevant for the evolution of the climate–ice sheet–carbon cycle system on timescales longer than thousands of years. Model parameters are calibrated using paleoclimate reconstructions and the results of two Earth system models of intermediate complexity. For a range of parameters values, the model is successful in reproducing the glacial–interglacial cycles of the last 800 kyr, with the best correlation between modelled and global paleo-ice volume of 0.86. Using different model realisations, we produce an assessment of possible trajectories for the next 1 million years under natural and several fossil-fuel CO2 release scenarios. In the natural scenario, the model assigns high probability of occurrence of long interglacials in the periods between the present and 120 kyr after present and between 400 and 500 kyr after present. The next glacial inception is most likely to occur ∼50 kyr after present with full glacial conditions developing ∼90 kyr after present. The model shows that even already achieved cumulative CO2 anthropogenic emissions (500 Pg C) are capable of affecting the climate evolution for up to half a million years, indicating that the beginning of the next glaciation is highly unlikely in the next 120 kyr. High cumulative anthropogenic CO2 emissions (3000 Pg C or higher), which could potentially be achieved in the next 2 to 3 centuries if humanity does not curb the usage of fossil fuels, will most likely provoke Northern Hemisphere landmass ice-free conditions throughout the next half a million years, postponing the natural occurrence of the next glacial inception to 600 kyr after present or later.