Global Environmental Change class content week 7 - 9 (Unit 3) Readings

Global surface temperatures are a broadly used indicator of

climate change and defined as an essential climate variable (ECV) by the Global Climate Observing System

There is not one single “surface

temperature” but rather a family of closely related but nonidentical temperatures: land surface air temperature (LSAT), land surface temperature, marine air temperature (MAT), sea surface temperature (SST), ice surface temperature, lake surface temperature

Typical historical global surface temperature change estimates are derived from combining

LSAT arising from fixed meteorological sites with SST estimates arising from ships and, more latterly, buoys

model-based projections generally use surface air temperature everywhere—equivalent to observed

LSAT and MAT

The basic LSAT data holdings contain myriad data artifacts that arise from factors as diverse as follows:

• •

  Station moves

• •

  Instrument changes

• •

  Observer changes

• •

  Automation

• •

  Time of observation biases

• •

  Microclimate exposure changes

• •

  Urbanization

There exist numerous national, regional, and global analyses that attempt to adjust for the nonclimatic artifacts—a process termed

homogenization

There are a variety of approaches to the challenge of homogenization. Early techniques tended to consider the stations in isolation or their characteristics relative to some composite of series from neighboring stations. Consideration of a station in isolation risks

misdiagnosing a real change in climate system behavior as a break in the series, thus adjusting away the real climate signal.

Several state-of-the-art homogenization techniques have been assessed against benchmark test cases [61,63]. Such test cases involve

presenting to the data set creators with data which have been synthetically produced and where the data originators know what the data issues required to be found and adjusted for are

Benchmarking exercises undertaken to date show that modern techniques tend to improve the

consistency and “correctness” of the records but that no technique is perfect.

The basic SST holdings have arisen from a broad range of measurement platforms using an array of measurement techniques that have changed substantially through time [33]. Biases in SST records are both larger and more systematic in nature than for

LSAT, and hence, homogenization is essential.

For SST, Measurements up until the 1940s were almost exclusively from

buckets whereby a sample of the sea water from just below the surface would be hauled onto the ship deck and measured

Since World War II, there has been a preponderance for either engine room intake–based measurement or hull contact sensors. Then since the 1990s, there has been an increasing ubiquity of drifting buoys so that today approximately

90% (by number but not coverage) of all measurements arise from this method.

In SST, Measurements based on buckets tended to be cold biased due to the effects of

evaporative cooling that occurs between sampling the water and its subsequent measurement.

Quite how cold biased depends upon the insulation efficiency of the bucket, the ship deck height, the delay between sampling and measurement, and the ambient weather conditions

The problem with SST and the bucket method

The evaporative cooling effect is greatest when windy and when the atmosphere is substantially warmer or colder than the sea surface. This is because

Without accounting for these effects, pre-1942 measurements would be too cold by c.0.3°C globally averaged.

Engine room intakes and hull contact sensors tend to sample water that has been

warmed relative to the ambient temperature by the ship itself and therefore be warm-biased.

Drifting buoys exhibit little obvious bias and substantially smaller spread than ship-based measurements. They measure temperatures that are about

0.12–0.18°C colder than the modern ships which are mainly engine room intake or hull sensor–based measurements

buoys with wooden to canvas bucket transition occurring between

1850 and 1920

from 1954 to 1975 an uncertain switch from

uninsulated to insulated buckets.

Global surface temperature data sets arise from combining

underlying data sets of LSAT and SST

The choice of whether to interpolate or not can have a significant impact, particularly on decadal

timescale behavior.

Sampling is not uniform in space or in time, and many regions have never been adequately sampled, these include

deserts, rainforests, polar regions, and regions of seasonal or perennial sea ice

If the temperatures in the unsampled regions are behaving in a way that is not represented in the remainder of the sampled portion of the globe, then a

biased estimate will result

It appears that over the past 20 years, in particular, interpolation has a distinct

effect upon apparent global mean behavior with interpolated analyses showing greater warming in the global mean

The change has not been linear in nature. There exist several decade-plus stretches of either little change or even cooling. This includes the early 21st century. This period had been dubbed a

“hiatus” and elicited much scientific and public interest, leading to its inclusion as a box in the IPCC Fifth Assessment Report

The suite of literature published on the matter strongly supported the assessment findings of Ref. [16] that the hiatus arose from a combination of

natural climate system variability and changes in short-lived, predominantly natural, climate system forcers

The Commission for Climatology Expert Team on Climate Change Detection and Indices (ETCDDI) has defined

27 core indices of which 16 are directly related to temperatures

Sea-level science is a rapidly developing field of climate science. Recent progress is summarized in two important reports:

the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC) [1] and the IPCC's Special Report on the Ocean and Cryosphere in a Changing Climate (SROCC)

One of the major advances highlighted in these reports is the ability of numerical climate models to

adequately simulate some dynamical ice sheet processes

Is sea level rising?

Although this question appears almost rhetorical and can at first glance be answered with a resounding “yes,” the direction of sea level-changes, positive or negative, depends on the

time scale of observations and the spatial scale under consideration.

However, sea-level changes are also highly variable in

time and space on decadal and longer timescales.

Satellite observations since the early 1990s have revealed the complex regional patterns of sea-level changes (Fig. 11.1A). Linear trends over the period 1993–present show that

global sea level is rising on average at a rate of about 3.4 mm/a

Some parts of the world's oceans have experienced rates of sea level rise that far exceed the global average (>10 mm/a in places), while in others,

sea level has fallen.

When averaged over the period 1955–2003 (Fig. 11.1B), linear rates are an order of magnitude smaller. The pattern, which is derived from tide gauges, satellite observations, and modeled reconstructions, is

Highly Complex

Many areas have been subjected to sea level rise, but there are places, most notably in the Indian Ocean and the tropical Pacific, where

sea level rise has been well below the global average, and in a few places, sea level has fallen, albeit by a small amount.

In coastal locations, land level movements need to be added to, or subtracted from, the mean sea-level change to

derive a figure that represents the relative change at a coastline. This is, after all, the number that is of most practical value from a coastal management perspective.

However, from the early 1990s to present, the rate of sea level rise was about

3.4 mm/a, with rates averaging 3.6 mm/a since the mid-2000s

Looking at sea level rise, while 20th century rates are based on tide gauges, more recent estimates are based on

both satellite and tide-gauge data, which are in reasonable agreement

According to the IPCC [2], the main contributors to sea level rise since 1993 have been

thermal expansion of the oceans (1.4 mm/a) and melting ice from small glaciers and ice caps (0.6 mm/a)

More recently, over the period 2006–15, the combined contribution of glaciers and ice sheets has reached

1.8 mm/a, exceeding the 1.4 mm/a contribution of thermal expansion over the same period

there has been a discrepancy between the contributions and observed sea level rise for the prealtimetry era. This has been called the

“sea level enigma” [18] or the “attribution problem”

“sea level enigma” [18] or the “attribution problem” implies that

one of the three things: either (1) the individual contributions have been underestimated by models, (2) there are sources of sea level rise that have not been accounted for, or (3) the measurements produced a global value that was too high.

One of the most uncertain terms in the sea level budget is the

contribution of terrestrial water sources.

Although the filling of reservoirs extracts water from the hydrological cycle and causes sea level to drop [20], other human interference with hydrological processes (e.g., wetland drainage, sedimentation in reservoirs, groundwater depletion, surface water consumption, deforestation) contribute positively to

Sea level rise

Volcanic eruptions have reduced the rate of

mean sea level rise, and some of the 20th-century rise in sea level was delayed by the eruptions of Krakatoa in 1886 and Pinatubo in 1991 [43], temporarily masking the impact of anthropogenic effects on sea level rise.

The relationship between sea-level change and greenhouse gas concentrations is well known on geological timescales. For example, when CO₂ concentrations were higher than 1000 ppm

around 70 Ma, ice was absent from the planet and sea level was 73 m higher than today

It is a well-known fact that rates of sea level rise in the past have been much higher than the ones we are experiencing today. For example, during the last deglaciation around 14000 years ago, rapid melting of ice sheets during Meltwater Pulse 1A produced rates of sea level rise in excess of

40 mm/a

The only remaining marine-based ice sheet is in West Antarctica, situated in one of the coldest regions of our planet. A comparison with the late glacial sea level history, therefore,

does not provide a suitable analog for modern (or future) conditions.

it is more instructive to examine periods in the Earth's history when the cryosphere contained roughly the same volume of ice as today (or slightly less), and temperatures were similar (or slightly higher) than today's temperatures. Periods often cited as useful analogs include the

Last Interglacial, the middle Holocene, and the Medieval Climatic Optimum.

Sea level rise is a major indicator of ongoing

Global change

Anthropogenic forcing by greenhouse gasses has become a dominant cause of

sea level rise, generating thermal expansion of ocean waters and melting of land-based ice.

Modern rates of sea level rise started about 100 years ago, and it is virtually certain that the 20th century rise is

faster than rates over the preceding three millennia.

However, during the Last Interglacial, rates of sea level rise were possibly

higher and may have been similar to those predicted in some future climate change scenarios.

The geological record of the past three glacial–interglacial cycles shows a strong positive relationship between

atmospheric CO₂ concentrations and sea level, and even if CO₂ remains at its current level, sea level rises in excess of several meters are likely over the next few centuries.

The increase of global mean temperatures due to anthropogenic greenhouse gas emissions especially in the past three decades, is leading to changes in

weather and climate extremes on local, regional, and global scales

Climate change not only impacts mean temperatures but also influences the whole

temperature distribution,

Shifting the distribution toward higher temperatures also shifts temperature extremes and leads to more

hot extremes and less cold extremes.

changes in the variability or shape of the temperature distribution can impact the

Extremes

Statistically speaking, extremes are rare events happening in the

tail of the distribution of a climate variable

A distinction between extreme weather and extreme climate events can be made, for instance, on the basis of their

temporal and spatial occurrence.

The timescale of extreme weather events typically ranges from

minutes to days, such as a storm or a heavy precipitation event,

an extreme climate event typically has a timescale of

months or years, such as a prolonged heatwave or drought

In the fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC), an extreme weather event was defined as

“an event that is rare at a particular place and time of year”

In the fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC), an extreme climate event as

“a pattern of extreme weather that persists for some time, such as a season”

For society, an important aspect of extreme events is their

potential for causing severe impacts if they occur at places where people and their assets are exposed to the extreme event and if the society

In the IPCC Special Report on Managing the Risk from Extreme Events [4], weather and climate extremes were assessed in the context of disaster risk reduction. The report emphasized that the hazards (i.e., weather and climate extremes) are not the sole determinants of risk, but that

the vulnerability and exposure of societies and natural systems play a significant part as well.

The IPCC [5] defines risk in the context of climate change as

“the potential for consequences where something of value is at stake and where the outcome is uncertain, recognizing the diversity of values. Risk is often represented as probability of occurrence of hazardous events or trends multiplied by the impacts if these events or trends occur.”

Tropical cyclones are a prominent example for extreme

weather causing severe impacts in specific regions

For analyzing extreme events with short timescales, such as heavy rainfall events that happen on a subdaily to daily

timescale, data are needed at relatively high temporal resolution (e.g., 10 min, hourly or 6-hourly).

to study changes over time, long-term (ideally at least five to six decades)

homogeneous data are required, covering as many regions of the world as possible.

Monthly gridded fields of temperature and precipitation observations are available over the past century [[10], [11], [12]], but daily or even subdaily data are much sparser, which makes it difficult to study

extremes on a global scale.

The Expert Team on Climate Change Detection and Indices (ETCCDI)¹ has developed a set of

27 indices based on daily maximum and minimum temperatures and daily precipitation that can be calculated for station-based observations (where available) and on gridded data sets, such as simulations of regional and global climate models

The ETCCDI indices mostly represent rather moderate aspects of climate extremes, for instance,

monthly or annual maxima and minima of temperature and precipitation, or the exceedance frequency of different relative or absolute thresholds

The increase of global mean temperatures throughout the recent decades has lead to corresponding changes in extreme temperatures. Observations of the past 40 years reveal increasing trends for both

minimum and maximum temperatures on major parts of the global land area

Annual maximum temperatures (TXx) show substantial and statistically significant increasing trends in large parts of

South America, Europe, and Asia.

annual minimum temperatures (TNn) exhibit large and statistically significant increasing trends in most

of South and North America, Africa, northern Eurasia, and the Tibetan Plateau.

The regions with strong trends in TXx and TNn differ considerably, which is due to different

processes influencing the evolution of maximum and minimum temperatures.

As an indicator of the hot season, TXx is closely connected to the radiation budget at

the Earth's surface

Daily maximum temperatures and thus TXx are affected by the

amount of incoming radiation and the partitioning of net radiation into latent and sensible heat fluxes.

Variations in cloud cover, increased

land drying, and (human-induced) land cover changes can have direct impacts on hot extremes

TNn, on the other hand, is an indicator of the

cold season and more sensitive to changes in snow cover and albedo.

(TXx) Means

Annual maximum temperatures

(TNn) Means

annual minimum temperatures

Extreme temperatures on land generally change at a faster rate than global mean temperature [26]. The reasons for this amplified increase of extreme temperatures are manifold and include the fact that

land areas warm up faster than the oceans (known as land-sea contrast,

This heterogeneity on regional scales reflects the important role of natural climate variability and the contribution of local processes and feedbacks, which can substantially impact the

effects of climate change at regional levels

Attributing extreme heat events to climate change requires a special modeling setup. For this purpose, climate model simulations with greenhouse gas concentrations at today's levels are compared with

simulations with greenhouse gas concentrations fixed at preindustrial levels. In other words, a world with climate change is compared with a hypothetical world where climate change did not happen.

Results from such studies consistently reveal a clear human influence on the historical evolution of extreme temperatures. For example, more than 75% of moderate and hot extremes under current climate conditions can be attributed to climate change

This is talking about comparing climate change to a world where climate change does not exist

The return time of extreme heat events was shown to have substantially decreased in recent years: Events happening twice a century in the early 2000s were expected to occur as often as twice a decade only about

15 years later

Heatwaves are often interlinked with drought conditions (see also Section 3), and importantly, they can mutually exacerbate each other. On the one hand, warmer air can hold more moisture, leading to enhanced

evapotranspiration rates and decreasing soil moisture

Warm air can carry more moisture than cool air. According to the Clausius–Clapeyron relationship, atmospheric water vapor content increases by

7% per degree Celsius, and observations indicate that extreme precipitation globally mostly follows this rate

In contrast, global mean precipitation is expected to increase at a lower rate of around

2%/°C

Where and how precipitation reaches the surface does not only depend on thermodynamics (i.e., the Clausius–Clapeyron relationship) but also on large-scale (e.g., frontal systems) and small-scale (e.g., summertime convection leading to thunderstorms)

atmospheric dynamics.

But what exactly is considered extreme precipitation? Unlike temperature, precipitation is not following a normal distribution, and it is

bounded by zero at the lower end.

However, like heatwaves, precipitation can be extreme in its intensity, frequency, and duration. Therefore,

various indices that capture different quantities of precipitation above a certain percentile or threshold and/or fall within a defined period of time are used in scientific studies.

Detecting trends for indices that capture extreme precipitation is, however,

more difficult than for temperature changes due to a comparatively low signal-to-noise ratio and less extensive coverage of observations

Extreme amounts of precipitation that fall in a region either because of high intensity, frequency (several events occurring close to each other), or precipitation falling during an extended period of time, as well as any combination of these three aspects, can lead to

Flooding

There are different types of floods, for example, flash floods or pluvial floods that are mostly caused by

heavy precipitation, floods following the passage of a tropical cyclone, or fluvial (river) floods

For the latter, the characteristics of the river catchments are key, in combination with the

preconditions of soil moisture, the time of year (e.g., large amounts of snow do not have such an immediate flooding impact as large amounts of precipitation might have), and how the river is managed, since currently, most rivers are not in their natural states.

However, a time period with less precipitation than usual can lead to a drought, in particular when combined with a deficit in

soil moisture or streamflow or increased atmospheric evaporative demand.

Some extreme events that have occurred in the past years and involve the hydrological cycle appear to have been made more likely by

anthropogenic forcing.