The acknowledgements section of scientific papers provides several pieces of information. This section identifies the funding sources that supported the work, as well as individuals who contributed to the research that didn't reach the level of being added as an author. This section also often provides details about the source of materials, data, information, and sometimes even ideas that are included in the paper. Here, the acknowledgements include links to the climate projections and reconstructions.

Taylor, Stouffer, and Meehl provided an overview of phase five of the Climate Model Intercomparison Project (CMIP5), written while the project was underway. The project aimed to produce a new set of freely available coordinated climate model experiments and data sets. The paper describes the plan for the experiments and a discussion of issues related to interpreting the results.

Indermühle and colleagues developed an atomospheric CO₂ concentration reconstruction for the entire Holocene from ice core samples from the Taylor Dome in Antarctica.

The ice cores have trapped air bubbles, which were only exposed to atmosphere when the bubbles were at the top (youngest) layer of the ice. The CO₂ concentration of these bubbles were measured using infared spectroscopy. By testing many layers throughout the entire core sample, the researchers reconstructed the atmospheric CO₂ levels throughout the last 11,000 years, to the start of the Holocene.

In this report, Yackulc and colleagues argue that niche-based models, which predict future species distribution using current plant or animal distributions, assume that the observed species are in equilibrium, which may not be a valid assumption for many ecosystems that are in flux. They proposed a process-based model that uses colonization and extinction rates for a given species in a given environment, and use case studies to show the model's applicability.

Cook and coauthors analyzed variations in droughts in the Mediterranean region over the past 900 years (1100–2012), using tree-ring data. They found a number of patterns in drought variation and distribution, and also determined a 98% likelihood that the current drought in the Levant region (Cyprus, Israel, Lebanon, Palestine, Syria, and Turkey) is the worst in the past 500 years, and an 89% likelihood that it is the worst in 900 years.

In this integrative study, Kaniewski, Van Campo, and Weiss examined the trends of water availability during the Medieval Climate Anomaly (900–1300 CE) and the Little Ice Age (1550–1850 CE) in Northern Mesopotamia, and how those trends affected human systems.

The study combined archaeological data, climate proxies, and an agricultural record based on pollen. The researchers found that changes in precipitation had an impact on yield and productivity of crops, and these changes were correlated with changes in agriculture and settlement activities.

Xoplaki and co-authors examined how different variables affected the wet season patterns across the Mediterranean region for the years 1950–1999. They found that while large-scale atmospheric features had a high influence on the variability, smaller scale factors also impacted the rainfall across the region.

Guito and Kaniewski present the "inverted" BIOME4 method of reconstructing climate variables and biomes from pollen core data that is the basis for the current paper.

They developed a climate data for a small set of gridpoints in the Mediterranean region for two time frames: 1901–2000 and 10,000 years ago to the present. They found that the climate dynamics of 1901–2000 differed from those of the past 10,000 years in several ways, and that the warming pattern in the past decades does not have an equivalent match in the earlier Holocene period.

The Intergovernmental Panel on Climate Change develops reports on climate change that provide information for the United Nations Framework Convention on Climate Change. The Fifth Assessment Report was published in 2014, in advance of the Paris Agreement meeting.

This report section focuses identifying impacts of climate change that have already happened. The report found more observations of regional climate change impacts than in the past, and reaching all continents, but also found gaps in knowledge of climate change impacts at regional levels.

The final map, 3H, uses a different color coding key, and shows the grid points that are different between the past, present, and future scenarios. Grid points that differ between the 4700 B.P. scenario and the present scenario are marked with blue circles. The filled dots all indicate differences in the biomes between the present and certain future scenarios. Yellow dots are grid points that are only different in the 4.0°C scenario, RCP8.5. The orange dots are different in both the RCP4.5 and RCP8.8 scenarios, and the red dots are different in the RCP2.6, RCP4.5, and RCP7.5 scenarios.

The third map, 3C, was generated by putting the current climate data into the BIOME4 model, running in "forward" mode. This does differ from the first map, based on pollen core samples. The biggest difference is that the first map shows warm mixed forest whereas the third map shows Temperate Deciduous Forest; this is likely caused by pine pollen being over-represented in the pollen core data.

The second map, 3B, comes from putting the 4700 B.P. pollen core data (covering the years 4650–4750 B.P.) into the BIOME4 model. This time slice was picked to represent the past Holocene because it had the highest variation from the present distribution - this is the bar that extends above the 99th percentile line in Figure 2.

Though this paper looks at all the plant ecosystems in the defined region, there are groups of plants that have been identified as characteristic of the Mediterranean climate, and biogeographers have identified regions where this mix of plants grow, generally distributed in the parts of the designated area that are closer to the Mediterranean Sea.

This specific mix of plants is not one of the biome types that is defined in the BIOME4 model, but the "warm mixed forest biome" shows a similar distribution, with two main exceptions.

Starting in 2030, the simulations for all four pathways start to differ in their biome change ratios. By 2100, the 1.5°C pathway, RCP2.6L, stays within the higher end of range of variation seen during the Holocene. The 2.0°C pathway, RCP2.6, has a level of change that is only seen in the single most extreme Holocene timeslice.

Both of the pathways with higher emissions have higher change ratios than those seen in any time in the Holocene.

This Pathway has the global emission peak between 2010–2020, and decrease substantially after that, with a reduction in the total concentration starting after 2050. This RCP results in a projection near the 2°C target.

Models of atmospheric greenhouse gas concentrations that were developed to provide researchers with a common set of pathways and their data for modeling and research. These pathways were used in the IPCC report published in 2014, in advance of the 2015 Paris Conference. The numbers refer to the increase in radiative forcing values The numbers refer to the increase in radiative forcing values (W/m²—a measure of how much more solar energy Earth is retaining because of the greenhouse effect) that would be expected in 2100 compared to pre-industrial levels.

These lines, marked below in purple, provide information about the distribution of the black bars, which show the proportional biome change in past Holocene compared to the current period. Half of the values are at or below the lowest line, whereas all but 1% of the past values fall at or below the top line.

Here's the figure with the lines more prominently marked:

For this figure, the years are not using the formal B.P. definition where "present" is 1950. In this case, "present" is 2000–2010.

To calculate future values, BIOME4 was run in its normal mode—using climate inputs to generate biome predictions. The climate inputs come from the models and time-series data found in the Coupled Model Intercomparison Project (CMIP5), which "promotes a standard set of model simulations." This way, research is done using the same underlying models, which allows for better comparisons across projects. Researchers update these models approximately every 5 years, incorporating new information.

The formal term Before Present (B.P.) refers to years before 1950, which was set as the dividing point during the early years of radiocarbon dating. When the term is not capitalized, that may mean it is being used informally and has a different time frame for the "present," such as the year when the paper is published or the most recent year of the reported data.

Were the authors using the term formally or informally? Given what you now know, what does the "present" mean in this description?

Most simulations show the Mediterranean Basin warming more than the global average.

Some climate variables and characteristics, such as temperature, leave no direct evidence of their values and variations in the historical record before humans began to observe and record them. However, researchers can use analysis and modeling of related data that are available, such as pollen cores, to build reconstructions of these characteristics and variables, and fill in gaps. In this case, a model uses pollen core data to reconstruct information about climate and the types of plant ecosystems in the Mediterranean Basin over the past 10,000 years.

The researchers use two main methods to generate the information about the Mediterranean Basin in this paper—for the past, the model uses data from pollen cores to generate information about the climate and ecosystems found over the past 10,000 years. For the future, the model uses different CO₂ emissions pathways to generate projections about the ecosystems and climate.

This data set was developed and maintained by the Climatic Research Unit in the United Kingdom. It provides several climate values for locations covering the globe, on a monthly basis, from 1901–2009. The data set covers the world's land surface (excluding Antarctica) and is reported in a 0.5° x 0.5° grid. The climate variables, including temperature and precipitation, are based on actual observations from a number of sources, using appropriate methods to integrate the different sources and calculate the values for locations that are missing data.

The error bars for the Holocene and 1901–2009 data points are based on standard deviations. This is a different method from what the researchers used to calculate the future predicted values.

How are local communities responding to reports like this one about the impacts of climate change in their region?

Read about how communities in the Maghreb region of North Africa are working to keep desert oasis ecosystems from disappearing at Yale Environment 360: http://e360.yale.edu/features/a_drive_to_save_sahran_oases_as_climate_change_takes_a_toll_cop22

The modeling presented in this paper indicates that limiting warming to a global average of 1.5°C above the pre-industrial average will result in changes to ecosystems; but, those changes will be within the range of variation that was seen in the past ~10,000 years, which includes most of human history.

However, at an average temperature increase at or above 2°C, the models indicate that the changes to the ecosystems will be far beyond what was seen in that time frame.

In the past, ecosystem variation was largely due to reductions in rainfall and other precipitation, as opposed to the temperature-induced changes that are expected if the global average temperature increases.

The Industrial Revolution led to a large increase in the amount of carbon dioxide released into the atmosphere as a result of human activities, so this is a common dividing point for looking at climate change. The Intergovernmental Panel on Climate Change sets this dividing line at the year 1750 for their reports.

The United Nations Framework Convention on Climate Change (UNFCCC) is an international treaty with the objective of "stabilization of greenhouse gas concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with the climate system." In other words, to limit human-produced greenhouse gases in order to stabilize Earth's global climate.

The treaty is a framework for negotiating future international treaties for action on those goals. It entered into force in 1994 after it was ratified by enough countries.

For the reconstructions of the past Holocene ecosystems based on the pollen cores, the initial model outputs are at a lower resolution than the 0.5° by 0.5° resolution of the future projected biomes. The resolution of the reconstructions was a grid of 2° latitude and 4° longitude.

To allow more direct comparisons between past and future, the researchers interpolated the reconstruction data to fill in a 0.5° by 0.5° grid. However, this may miss some features at scales smaller than the original gird.

Data or information that is specific to both a time and location. In this case, the "spacio-" aspect is the set of 0.5° x 0.5° grid points across the Mediterranean land mass. The "-temporal" aspect is the specific time period examined— either in the past or future.

Schleussner and co-authors assessed the impacts of reaching the 1.5°C and 2.0°C thresholds for a number of variables, at a regional level. They found large differences in the impacts of the two scenarios, covering a range of issues from the fraction of global coral reefs at risk of annual bleaching to heat wave duration.

In most scientific journals the Materials and Methods are part of the main text, but in Science they are part of the references and Supplementary Materials. For details on how the researchers developed their pathways and models, and information on the different ecosystem groupings, check out that link.

Prior work from author Guiot developed a reconstruction of the Mediterranean region's climate over the past 10,000 years. The reconstruction used the BIOME4 model, with pollen core data as the input. The reconstruction found that the past climate variations were dominated by changes in precipitation (droughts and wet periods). However, unlike in the present, the drought periods were linked to cool temperatures. They also found that the pattern of warming was different in the past than it is currently.

Using this information, they concluded that the reconstruction indicates that recent warming in the region is unlike what was seen during any other time in the Holocene.

These are some of the main general biomes found in the Mediterranean region. The BIOME4 model uses a longer list of specific biome types, and more than 20 were found in the Mediterranean region. For the analysis, this larger list was grouped into ten “aggregated biome types” which are presented in Figure 3. To see how the aggregated biomes were grouped, check out Table S1 in the Supplementary Materials.

BIOME4 was developed by Kaplan and colleagues. The model uses input data about water, climate, and soil to simulate the ecosystem and plant processes that drive the distribution of vegetation types. Using this information, the model identifies key variables and uses these to rank the likelihood of different groupings of plants for the given time and location. The rankings then help identify the biome type present.

This model predicts plant distribution based on underlying processes that are expected to be consistent for the entire Holocene time period–for example, net assimilation of CO₂ (the flux of CO₂ between a leaf and the atmosphere) or the water balance in an ecosystem. The authors contrast that method to other models that are tuned based on how well the results correlate with, or reproduce, current vegetation patterns.

Each of the colored lines is generated from using multiple climate models as input to the BIOME4 model. The solid line indicates the average change ratio predicted from multiple runs of the model.The variation within each model is indicated by the shaded region, which encompasses the 25% of results above and below the average. As you can see, some models had higher variability than others, and the predicted changes within the first 50 years overlap between models.

Shrubs and plants that have adapted to survive in locations with very low amounts of liquid water, including deserts as well as regions with ice and snow.

Though the BIOME4 methods have limitations, there are two major advantages.

First, the model is based on the underlying processes that connect climate and ecosystems, so it avoids the assumptions made by models that are based on correlating current ecosystem distributions and climate values. For example, these models often assume that the current ecosystem distributions are stable and at equilibrium, rather than in flux.

Second, the researchers are able to use the same model for both the past reconstructions and for the future projections. This allows for more direct comparison between the two groups.

To reconstruct Holocene climate variables and biome types, the researchers "inverted" the BIOME4 model, flipping the input and output data. Rather than using climate data as the input, it uses plant data.

The researchers started by converting the pollen core data for a given time and location into "plant functional type" scores, which BIOME4 uses to rank and select biome types based on the "best match." "Plant functional type" groups species together based on similar characteristics such as respiration rate, response to climate variation, and genetic makeup.

The researchers then ran the BIOME4 model in "inverse" mode, testing a large number of randomized climate inputs for each time and location. The set of climate inputs that resulted in the best match to the pollen core data is used for the reconstruction.

Kelly and colleagues examined the 3-year drought in Syria that began in Winter 2006/2007 to understand what contributed to the severity of the drought, which was the worst in human record.

The researchers found that though multiyear droughts do occur naturally in the region, the reduced precipitation and increased surface temperature of the past century that is consistent with models of human-driven climate impacts contributed to the extreme nature of the drought.

The impacts of the drought were also amplified by previous droughts, the reduction of groundwater supply (which had previously acted as a buffer), and population increases in recent years.

The current period is experiencing both reduced precipitation and high temperatures, unlike during the earlier Holocene, when times of low precipitation were generally cooler.

Precipitation and temperatures vary across the Mediterranean region, and the precipitation patterns are more variable than the temperature patterns. In general, the changes of these two climate variables in the next few decades are expected to lead to a reduction in the amount of water available as well as the loss of some ecosystems and biodiversity.

In many scientific journals, the Materials and Methods are part of the main text, but in Science they are part of the references and Supplementary Materials. For details on how the researchers developed their pathways and models, and information on the different ecosystem groupings, check out the supplementary information.

Guilot, one of the authors of the current paper, and his collaborator used pollen core data and vegetation modeling to develop reconstructions of the Mediterranean Basin climate over the past 10,000 years. To do this, they ran BIOME4, a vegetation model, in "inverse" mode.

BIOME4 is typically used to predict vegetation by inputting climate information. In an "inverse" mode, researchers input the vegetation information from the pollen cores and then used a computer algorithm to calculate what climate would have produced the mix of plants that was observed.

This reconstruction method is the basis of the current paper.

These boxes show a range of results from many model runs for each scenario. The dot indicates the 50th percentile (median) result for all runs. The box shows the range containing the middle 50% of the results.

People receive a wide range of benefits from the environment, and these benefits are sometimes referred to as "ecosystem services."

The lowercase of the Greek letter sigma is used to indicate standard deviation. Here's a short video explanation: https://www.youtube.com/embed/MRqtXL2WX2M

These scenarios are different models of atmospheric greenhouse gas concentrations. Descriptions of each model can be found in a special issue of Climate Change from November 2011: https://link.springer.com/journal/10584/109/1/page/1

The targets are for global temperatures, based on averages across the entire surface of the planet. Because of climate variations, this does not mean that each location would experience the same degree of local climate impacts.

For example, researchers examined the impact of the global 2°C increase scenario on regional changes to temperature extremes and their analysis found regions that would have a bigger response. For the Mediterranean region, they predicted a 3°C increase in hot temperature extremes.

The Paris Agreement includes specific actions like the Nationally Determined Contributions (NDC) each country identifies and reports; these include target national emission rates over time, and whether these have been achieved. The Paris Agreement has been ratified, but related work continues.

For more information on the details of the Paris Agreement, read more in the UNFCCC's Paris Agreement Hub: http://unfccc.int/paris_agreement/items/9485.php

For updates about the Paris Agreement, read more in the UNFCCC's Paris Agreement Newsroom: http://newsroom.unfccc.int/paris-agreement/

An agreement under the UNFCCC framework addressing three main aims:

• lessening greenhouse gas emissions,
• preparing for and negating the effects of climate change, and
• financing dedicated to accomplish the previous two aims.

Under the agreement, countries set targets to lessen global warming, develop plans to meet those targets, and report on their progress.

The agreement was negotiated in Paris at the 21st Conference of the Parties of the UNFCCC in December 2015 and adopted in 2016.

Pollen grains are made of materials that are highly resistant to breaking down in the environment, so they can be found in sediments going back thousands of years (and even further in the fossil record). The grains are diverse and researchers can identify the plant types they come from by examining them under a microscope. This allows researchers to see which seed-baring plants were in a given location in the past by taking sediment core samples and examining the pollen grains in each layer. The deeper the layer, the older the layer.

The geographic area surrounding the Mediterranean Sea. There are many different sets of boundaries for this region used by researchers, historians, and others. For this paper, the authors use the region between longitudes 10°W to 45°E and latitudes 28°N to 48°N. The region can be seen in Figure 3, and an approximate map of the region is included here.

Mountain ecosystems.

The authors do not include changes that occur in the RCP2.6L scenario in the Figure 3H map. Why might they have left that out?

The RCP2.6 scenario, the pathway that is likely to limit warming to the 2.0°C threshold targeted by the Paris Agreement, would result in changes that are just within the most extreme variation of the Holocene. This pathway is considered highly ambitious as it would require achieving even more emissions reductions than are currently proposed.

The pathway developed by the authors to meet the 1.5° threshold, RCP2.6L, is the only one that would keep the ecosystem changes within the Holocene variation.

This analysis indicates that for the Mediterranean region, there will likely be significantly larger impacts if the global temperature increase reaches the threshold of 2.0°C rather than being limited to 1.5°C.

The current scenario for global greenhouse gas emissions, taking into account all the voluntary emission reductions targets set by countries as part of the United Nations Framework Convention on Climate Change, matches best with RCP4.5. This analysis shows that the climate change associated with that pathway would likely result in changes in the Mediterranean ecosystems by the end of the century that go far beyond what was seen in the past 10,000 years.

Plants with hard leaves that are close together on the stem, adapted to hot and dry conditions.

Farming is one type of land use that is both heavily impacted by and contributes to changes in the environment. Learn more about how drought conditions in Spain are affecting and affected by farming in NPR Parallels: http://www.npr.org/sections/parallels/2017/07/02/534746211/drought-threatens-crops-wildlife-along-spains-guadalquivir-river-delta?ft=nprml&f=%20NPR

Even though the Atlantic pine forests and the Mediterranean vegetation types are different, the BIOME4 model is not able to sort the two types into separate categories.

Plains or grasslands, generally without trees except in areas around water like lakes or rivers.

A 2°C change is about the same as a 3.6°F change.

The four maps for the future pathways, 3D to 3G, are based on putting the climate data for year 2100 for each pathway into the BIOME4 model, running in "forward" mode.

The first map, 3A, comes from putting pollen core data from the past century into the BIOME4 inversion model, the same method as is used for the past biome map (3B).

The past 10,000 years of the Holocene includes all of recorded human history, and is a common timeframe for examining the recent (geologically speaking) past. The geological epoch of the Holocene began about 11,700 years ago with the end of the last major ice age.

This RCP assumes that greenhouse gas emissions continue to rise from now through 2100. The Materials and Methods section calls this the "business as usual" scenario, and indicates this would result in an average global temperature increase of 4°C by 2100.

None of the existing RCPs resulted in a projection of 1.5°C, so the researchers created a new model so that they could evaluate the scenario.

In this RCP, the greenhouse gas emissions increase until 2040, and then decrease. This results in the total concentration leveling out after approximately 2060.