Editing IPCC:AR6/SRCCL/Chapter-2 (section)

== 2.2 The effect of climate variability and change on land ==

<span id="overview-of-climate-impacts-on-land"></span>
=== 2.2.1 Overview of climate impacts on land ===

<div id="section-2-2-1-1-climate-drivers-of-land-form-and-function"></div>

<span id="climate-drivers-of-land-form-and-function"></span>
==== 2.2.1.1 Climate drivers of land form and function ====

<div id="section-2-2-1-1-climate-drivers-of-land-form-and-function-block-1"></div>

Energy is redistributed from the warm equator to the colder poles through large-scale atmospheric and oceanic processes driving the Earth’s weather and climate (Oort and Peixóto 1983 <sup>[[#fn:r26|26]]</sup> ; Carissimo et al. 1985 <sup>[[#fn:r27|27]]</sup> ; Yang et al. 2015a <sup>[[#fn:r28|28]]</sup> ). Subsequently, a number of global climate zones have been classified ranging from large-scale primary climate zones (tropical, sub-tropical, temperate, sub-polar, polar) to much higher-resolution, regional climate zones (e.g., the Köppen-Geiger classification, Kottek et al. 2006 <sup>[[#fn:r29|29]]</sup> ). Biomes are adapted to regional climates (Figure 2.1) and may shift as climate, land surface characteristics (e.g., geomorphology, hydrology), CO <sub>2</sub> fertilisation and fire interact. These biomes and the processes therein are subject to modes of natural variability in the ocean-atmosphere system that result in regionally wetter/dryer or hotter/cooler periods having temporal scales from weeks to months (e.g., Southern Annular Mode), months to seasons (e.g., Madden-Julian Oscillation), years (e.g., El Niño Southern Oscillation) and decades (e.g., Pacific Decadal Oscillation). Furthermore, climate and weather extremes (such as drought, heatwaves, very heavy rainfall, strong winds), whose frequency, intensity and duration are often a function of large-scale modes of variability, impact ecosystems at various space and timescales.

It is ''very likely'' that changes to natural climate variability as a result of global warming has and will continue to impact terrestrial ecosystems with subsequent impacts on land processes (Hulme et al. 1999 <sup>[[#fn:r30|30]]</sup> ; Parmesan and Yohe 2003 <sup>[[#fn:r31|31]]</sup> ; Di Lorenzo et al. 2008 <sup>[[#fn:r32|32]]</sup> ; Kløve et al. 2014 <sup>[[#fn:r33|33]]</sup> ; Berg et al. 2015 <sup>[[#fn:r34|34]]</sup> ; Lemordant et al. 2016 <sup>[[#fn:r35|35]]</sup> ; Pecl et al. 2017 <sup>[[#fn:r36|36]]</sup> ). This chapter assesses climate variability and change, particularly extreme weather and climate, in the context of desertification, land degradation, food security and terrestrial ecosystems more generally. This section does specifically assess the impacts of climate variability and climate change on desertification, land degradation and food security as these impacts are assessed respectively in Chapters 3, 4 and 5. This chapter begins with an assessment of observed warming on land.

<div id="section-2-2-1-1-climate-drivers-of-land-form-and-function-block-2"></div>

<span id="figure-2.1"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 2.1'''

<span id="worldwide-bioclimatic-classification-system-19962018.source-rivas-martinez-et-al.-2011.-online-at-www.globalbioclimatics.org"></span>
<!-- IMG CAPTION -->
'''Worldwide Bioclimatic Classification System, 1996–2018.Source: Rivas-Martinez et al. (2011). Online at www.globalbioclimatics.org'''
<!-- IMG FILE -->
[[File:6bb2aee1284046e6e1b0ccde2455bf3e Figure-2.1-1024x504.png]]

Worldwide Bioclimatic Classification System, 1996–2018.Source: Rivas-Martinez et al. (2011). Online at [[IPCC:Srccl:Chapter:Chapter-2:Www.globalbioclimatics.org|www.globalbioclimatics.org]]

<!-- END IMG -->
<div id="section-2-2-1-2-changes-in-global-land-surface-air-temperature"></div>

<span id="changes-in-global-land-surface-air-temperature"></span>

==== 2.2.1.2 Changes in global land surface air temperature ====

<div id="section-2-2-1-2-changes-in-global-land-surface-air-temperature-block-1"></div>

Based on analysis of several global and regional land surface air temperature (LSAT) datasets, AR5 concluded that the global LSAT had increased over the instrumental period of record, with the warming rate approximately double that reported over the oceans since 1979 and that ‘it is certain that globally averaged LSAT has risen since the late 19th century and that this warming has been particularly marked since the 1970s’. Warming found in the global land datasets is also in a broad agreement with station observations (Hartmann et al. 2013a <sup>[[#fn:r37|37]]</sup> ).

Since AR5, LSAT datasets have been improved and extended. The National Center for Environmental Information, which is a part of the US National Oceanic and Atmospheric Administration (NOAA), developed a new, fourth version of the Global Historical Climatology Network monthly dataset (GHCNm, v4). The dataset provides an expanded set of station temperature records with more than 25,000 total monthly temperature stations compared to 7200 in versions v2 and v3 (Menne et al. 2018 <sup>[[#fn:r38|38]]</sup> ). Goddard Institute for Space Studies, which is a part of the US National Aeronautics and Space Administration, (NASA/ GISS), provides estimate of land and ocean temperature anomalies (GISTEMP). The GISTEMP land temperature anomalies are based upon primarily NOAA/GHCN version 3 dataset (Lawrimore et al. 2011 <sup>[[#fn:r39|39]]</sup> ) and account for urban effects through nightlight adjustments (Hansen et al. 2010 <sup>[[#fn:r40|40]]</sup> ). The Climatic Research Unit (CRU) of the University of East Anglia, UK (CRUTEM) dataset, now version CRUTEM4.6,

incorporates additional stations (Jones et al. 2012 <sup>[[#fn:r41|41]]</sup> ). Finally, the Berkeley Earth Surface Temperature (BEST) dataset provides LSAT from 1750 to present based on almost 46,000 time series and has the longest temporal coverage of the four datasets (Rohde et al. 2013 <sup>[[#fn:r42|42]]</sup> ). This dataset was derived with methods distinct from those used for development of the NOAA GHCNm, NASA/GISS GISTEMP and the University of East Anglia CRUTEM datasets.

<div id="section-2-2-1-2-changes-in-global-land-surface-air-temperature-block-2"></div>

<span id="table-2.1"></span>
<!-- START TABLE -->
'''Table 2.1'''

<span id="increases-in-land-surface-air-temperature-lsat-from-preindustrial-period-and-the-late-19th-century-to-present-day."></span>
'''Increases in land surface air temperature (LSAT) from preindustrial period and the late 19th century to present day.'''

<!-- TABLE -->
{| class="wikitable"
|-
|

Dataset of LSAT increase (°C)

|-

Time period

BEST

CRUTEM4.6

GHCNm, v4

GISTEMP

|-

''From'' 1850–1900 to 2006–2015

1.53<br />
1.38–1.68<br />
(95% confidence)

1.32*

|
|-

''From'' 1850–1900 to 1999–2018

1.52<br />
1.39–1.66<br />
(95% confidence)

1.31

NA

NA

|-

''From'' 1881–1900 to 1999–2018

1.51<br />
1.40–1.63<br />
(95% confidence)

1.31

1.37

1.45

|}
<!-- END TABLE -->

\* CRUTEM4.6 LSAT increase is computed from 1856–1900 average.

<div id="section-2-2-1-2-changes-in-global-land-surface-air-temperature-block-3"></div>

According to the available observations in the four datasets, the globally averaged LSAT increased by 1.44°C from the preindustrial period (1850–1900) to the present day (1999–2018).The warming from the late 19th century (1881–1900) to the present day (1999–2018) was 1.41°C (1.31°C–1.51°C) (Table 2.1). The 1.31°C–1.51°C range represents the spread in median estimates from the four available land datasets and does not reflect uncertainty in data coverage or methods used. Based on the BEST dataset (the longest dataset with the most extensive land coverage) the total observed increase in LSAT between the average of the 1850–1900 period and the 2006– 2015 period was 1.53°C, (1.38–1.68°C; 95% confidence), while the GMST increase for the same period was 0.87°C (0.75–0.99°C; 90% confidence) (IPCC, 2018: Summary for policymakers, Allen et al. 2018 <sup>[[#fn:r43|43]]</sup> ).

The extended and improved land datasets reaffirmed the AR5 conclusion that it is certain that globally averaged LSAT has risen since the preindustrial period and that this warming has been particularly marked since the 1970s (Figure 2.2).

Recent analyses of LSAT and sea surface temperature (SST) observations, as well as analyses of climate model simulations, have refined our understanding of underlying mechanisms responsible for a faster rate of warming over land than over oceans. Analyses of paleo records, historical observations, model simulations and underlying physical principles are all in agreement that the land is warming faster than the oceans as a result of differences in evaporation, land–climate feedbacks (Section 2.5) and changes in the aerosol forcing over land ( ''very high confidence'' ) (Braconnot et al. 2012 <sup>[[#fn:r44|44]]</sup> ; Joshi et al. 2013 <sup>[[#fn:r45|45]]</sup> ; Sejas et al. 2014 <sup>[[#fn:r46|46]]</sup> ; Byrne and O’Gorman 2013 <sup>[[#fn:r47|47]]</sup> , 2015 <sup>[[#fn:r48|48]]</sup> ; Wallace and Joshi 2018 <sup>[[#fn:r49|49]]</sup> ; Allen et al. 2019 <sup>[[#fn:r50|50]]</sup> ). There is also ''high confidence'' that difference in land and ocean heat capacity is not the primary reason for faster land than ocean warming. For the recent period, the land-to-ocean warming ratio is in close agreement between different observational records (about 1.6) and the CMIP5 climate model simulations (the ''likely'' range of 1.54°C to 1.81°C). Earlier studies analysing slab ocean models (models in which it is assumed that the deep ocean has equilibrated) produced a higher land temperature increases than sea surface temperature (Manabe et al. 1991 <sup>[[#fn:r51|51]]</sup> ; Sutton et al. 2007 <sup>[[#fn:r52|52]]</sup> ).

It is certain that globally averaged LSAT has risen faster than GMST from the preindustrial period (1850–1900) to the present day (1999–2018). This is because the warming rate of the land compared to the ocean is substantially higher over the historical period (by approximately 60%) and because the Earth’s surface is approximately one-third land and two-thirds ocean. This enhanced land warming impacts land processes with implications for desertification (Section 2.2.2 and Chapter 3), food security (Section 2.2.3 and Chapter 5), terrestrial ecosystems (Section 2.2.4), and GHG and non-GHG fluxes between the land and climate (Sections 2.3 and 2.4). Future changes in land characteristics through adaptation and mitigation processes and associated land–climate feedbacks can dampen warming in some regions and enhance warming in others (Section 2.5).

<div id="section-2-2-1-2-changes-in-global-land-surface-air-temperature-block-4"></div>

<span id="figure-2.2"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 2.2'''

<span id="evolution-of-land-surface-air-temperature-lsat-and-global-mean-surface-temperature-gmst-over-the-period-of-instrumental-observations.-the-brown-line-shows-annual-mean-lsat-in-the-best-crutem4.6-ghcnmv4-and-gistemp-datasets-expressed-as-departures-from-global-average-lsat-in-18501900-with-the-brown-line-thickness-indicating-inter-dataset-range.-the-blue-line-shows"></span>
<!-- IMG CAPTION -->
'''Evolution of land surface air temperature (LSAT) and global mean surface temperature (GMST) over the period of instrumental observations. The brown line shows annual mean LSAT in the BEST, CRUTEM4.6, GHCNmv4 and GISTEMP datasets, expressed as departures from global average LSAT in 1850–1900, with the brown line thickness indicating inter-dataset range. The blue line shows […]'''
<!-- IMG FILE -->
[[File:e3db74bab240f8ac71cfee37c41f7fd2 Figure-2.2-1024x670.jpg]]

Evolution of land surface air temperature (LSAT) and global mean surface temperature (GMST) over the period of instrumental observations. The brown line shows annual mean LSAT in the BEST, CRUTEM4.6, GHCNmv4 and GISTEMP datasets, expressed as departures from global average LSAT in 1850–1900, with the brown line thickness indicating inter-dataset range. The blue line shows annual mean GMST in the HadCRUT4, NOAAGlobal Temp, GISTEMP and Cowtan&amp;Way datasets (monthly values of which were reported in the Special Report on Global Warming of 1.5ºC; Allen et al. 2018 <sup>[[#fn:r53|53]]</sup> ).

<!-- END IMG -->
<span id="climate-driven-changes-in-aridity"></span>
=== 2.2.2 Climate-driven changes in aridity ===

<div id="section-2-2-2-climate-driven-changes-in-aridity-block-1"></div>

Desertification is defined and discussed at length in Chapter 3 and is a function of both human activity and climate variability and change. There are uncertainties in distinguishing between historical climate- caused aridification and desertification and future projections of aridity as different measurement methods of aridity do not agree on historical or projected changes (Sections 3.1.1 and 3.2.1). However, warming trends over drylands are twice the global average (Lickley and Solomon 2018 <sup>[[#fn:r54|54]]</sup> ) and some temperate drylands are projected to convert to subtropical drylands as a result of an increased drought frequency causing reduced soil moisture availability in the growing season (Engelbrecht et al. 2015 <sup>[[#fn:r55|55]]</sup> ; Schlaepfer et al. 2017 <sup>[[#fn:r56|56]]</sup> ). We therefore assess with ''medium confidence'' that a warming climate will result in regional increases in the spatial extent of drylands under mid- and high emission scenarios and that these regions will warm faster than the global average warming rate.

<span id="the-influence-of-climate-change-on-food-security"></span>
=== 2.2.3 The influence of climate change on food security ===

<div id="section-2-2-3-the-influence-of-climate-change-on-food-security-block-1"></div>

Food security and the various components thereof are addressed in depth in Chapter 5. Climate variables relevant to food security and food systems are predominantly temperature and precipitation-related, but also include integrated metrics that combine these and other variables (e.g., solar radiation, wind, humidity) and extreme weather and climate events including storm surge (Section 5.2.1). The impact of climate change through changes in these variables is projected to negatively impact all aspects of food security (food availability, access, utilisation and stability), leading to complex impacts on global food security ( ''high confidence'' ) (Chapter 5, Table 5.1).

Climate change will have regionally distributed impacts, even under aggressive mitigation scenarios (Howden et al. 2007 <sup>[[#fn:r57|57]]</sup> ; Rosenzweig et al. 2013 <sup>[[#fn:r58|58]]</sup> ; Challinor et al. 2014 <sup>[[#fn:r59|59]]</sup> ; Parry et al. 2005 <sup>[[#fn:r60|60]]</sup> ; Lobell and Tebaldi 2014 <sup>[[#fn:r61|61]]</sup> ; Wheeler and Von Braun 2013 <sup>[[#fn:r62|62]]</sup> ). For example, in the northern hemisphere the northward expansion of warmer temperatures in the middle and higher latitudes will lengthen the growing season (Gregory and Marshall 2012 <sup>[[#fn:r63|63]]</sup> ; Yang et al. 2015b <sup>[[#fn:r64|64]]</sup> ) which may benefit crop productivity (Parry et al. 2004 <sup>[[#fn:r65|65]]</sup> ; Rosenzweig et al., 2014 <sup>[[#fn:r66|66]]</sup> ; Deryng et al. 2016 <sup>[[#fn:r67|67]]</sup> ). However, continued rising temperatures are expected to impact global wheat yields by about 4–6% reductions for every degree of temperature rise (Liu et al. 2016a <sup>[[#fn:r68|68]]</sup> ; Asseng et al. 2015 <sup>[[#fn:r69|69]]</sup> ) and across both mid- and low latitude regions, rising temperatures are also expected to be a constraining factor for maize productivity by the end of the century (Bassu et al. 2014 <sup>[[#fn:r70|70]]</sup> ; Zhao et al. 2017 <sup>[[#fn:r71|71]]</sup> ). Although there has been a general reduction in frost occurrence during winter and spring, and a lengthening of the frost free season in response to growing concentrations of GHGs (Fischer and Knutti 2014 <sup>[[#fn:r72|72]]</sup> ; Wypych et al. 2017 <sup>[[#fn:r73|73]]</sup> ), there are regions where the frost season length has increased, for example, in southern Australia (Crimp et al. 2016 <sup>[[#fn:r74|74]]</sup> ). Despite the general reduced frost season length, late spring frosts may increase risk of damage to warming induced precocious vegetation growth and flowering (Meier et al. 2018 <sup>[[#fn:r75|75]]</sup> ). Observed and projected warmer minimum temperatures have, and will continue to, reduce the number of winter chill units required by temperate fruit and nut trees (Luedeling 2012 <sup>[[#fn:r76|76]]</sup> ). Crop yields are impacted negatively by increases of seasonal rainfall variability in the tropics, sub-tropics, water-limited and high elevation environments, while drought severity and growing season temperatures also have a negative impact on crop yield (Nelson et al. 2009 <sup>[[#fn:r77|77]]</sup> ; Schlenker and Lobell 2010 <sup>[[#fn:r78|78]]</sup> ; Müller et al. 2017 <sup>[[#fn:r79|79]]</sup> ; Parry et al. 2004 <sup>[[#fn:r80|80]]</sup> ; Wheeler and Von Braun 2013 <sup>[[#fn:r81|81]]</sup> ; Challinor et al. 2014 <sup>[[#fn:r82|82]]</sup> ).

Changes in extreme weather and climate (Section 2.2.5) have negative impacts on food security through regional reductions of crop yields. A recent study shows that 18–43% of the explained yield variance of four crops (maize, soybeans, rice and spring wheat) is attributable to extremes of temperature and rainfall, depending on the crop type (Vogel et al. 2019 <sup>[[#fn:r83|83]]</sup> ). Climate shocks, particularly severe drought impact low-income small-holder producers disproportionately (Vermeulen et al. 2012 <sup>[[#fn:r84|84]]</sup> ; Rivera Ferre 2014 <sup>[[#fn:r85|85]]</sup> ). Extremes also compromise critical food supply chain infrastructure, making transport of and access to harvested food more difficult (Brown et al. 2015 <sup>[[#fn:r86|86]]</sup> ; Fanzo et al. 2018 <sup>[[#fn:r87|87]]</sup> ). There is ''high confidence'' that the impacts of enhanced climate extremes, together with non-climate factors such as nutrient limitation, soil health and competitive plant species, generally outweighs the regionally positive impacts of warming (Lobell et al. 2011 <sup>[[#fn:r88|88]]</sup> ; Leakey et al. 2012 <sup>[[#fn:r89|89]]</sup> ; Porter et al. 2014 <sup>[[#fn:r90|90]]</sup> ; Gray et al. 2016 <sup>[[#fn:r91|91]]</sup> ; Pugh et al. 2016 <sup>[[#fn:r92|92]]</sup> ; Wheeler and Von Braun 2013 <sup>[[#fn:r93|93]]</sup> ; Beer 2018 <sup>[[#fn:r94|94]]</sup> ).

<span id="climate-driven-changes-in-terrestrial-ecosystems"></span>
=== 2.2.4 Climate-driven changes in terrestrial ecosystems ===

<div id="section-2-2-4-climate-driven-changes-in-terrestrial-ecosystems-block-1"></div>

Previously, the IPCC AR5 reported ''high confidence'' that the Earth’s biota composition and ecosystem processes have been strongly affected by past changes in global climate and that the magnitudes of projected changes for the 21st century under high warming scenarios (for example, RCP8.5) are higher than those under historic climate change (Settele et al. 2014 <sup>[[#fn:r95|95]]</sup> ). There is ''high confidence'' that as a

result of climate changes over recent decades many plant and animal species have experienced range size and location changes, altered abundances and shifts in seasonal activities (Urban 2015 <sup>[[#fn:r96|96]]</sup> ; Ernakovich et al. 2014 <sup>[[#fn:r97|97]]</sup> ; Elsen and Tingley 2015 <sup>[[#fn:r98|98]]</sup> ; Hatfield and Prueger 2015 <sup>[[#fn:r99|99]]</sup> ; Savage and Vellend 2015 <sup>[[#fn:r100|100]]</sup> ; Yin et al. 2016 <sup>[[#fn:r101|101]]</sup> ; Pecl et al. 2017 <sup>[[#fn:r102|102]]</sup> ; Gonsamo et al. 2017 <sup>[[#fn:r103|103]]</sup> ; Fadrique et al. 2018 <sup>[[#fn:r104|104]]</sup> ; Laurance et al. 2018 <sup>[[#fn:r105|105]]</sup> ). There is ''high confidence'' that climate zones have already shifted in many parts of the world, primarily as an increase of dry, arid climates accompanied by a decrease of polar climates (Chan and Wu 2015 <sup>[[#fn:r106|106]]</sup> ; Chen and Chen 2013 <sup>[[#fn:r107|107]]</sup> ; Spinoni et al. 2015b <sup>[[#fn:r108|108]]</sup> ). Regional climate zones shifts have been observed over the Asian monsoon region (Son and Bae 2015 <sup>[[#fn:r109|109]]</sup> ), Europe (Jylhä et al. 2010 <sup>[[#fn:r110|110]]</sup> ), China (Yin et al. 2019 <sup>[[#fn:r111|111]]</sup> ), Pakistan (Adnan et al. 2017 <sup>[[#fn:r112|112]]</sup> ), the Alps (Rubel et al. 2017 <sup>[[#fn:r113|113]]</sup> ) and north-eastern Brazil, southern Argentina, the Sahel, Zambia and Zimbabwe, the Mediterranean area, Alaska, Canada and north-eastern Russia (Spinoni et al. 2015b <sup>[[#fn:r114|114]]</sup> ).

There is ''high confidence'' that bioclimates zones will further shift as the climate warms (Williams et al. 2007 <sup>[[#fn:r115|115]]</sup> ; Rubel and Kottek 2010 <sup>[[#fn:r116|116]]</sup> ; Garcia et al. 2016 <sup>[[#fn:r117|117]]</sup> ; Mahony et al. 2017 <sup>[[#fn:r118|118]]</sup> ; Law et al. 2018 <sup>[[#fn:r120|120]]</sup> ). There is also ''high confidence'' that novel, unprecedented climates (climate conditions with no analogue in the observational record) will emerge, particularly in the tropics (Williams and Jackson 2007 <sup>[[#fn:r121|121]]</sup> ; Colwell et al. 2008a <sup>[[#fn:r122|122]]</sup> ; Mora et al. 2013 <sup>[[#fn:r123|123]]</sup> , 2014 <sup>[[#fn:r124|124]]</sup> ; Hawkins et al. 2014 <sup>[[#fn:r126|126]]</sup> ; Mahony et al. 2017 <sup>[[#fn:r127|127]]</sup> ; Maule et al. 2017 <sup>[[#fn:r128|128]]</sup> ). It is ''very likely'' that terrestrial ecosystems and land processes will be exposed to disturbances beyond the range of current natural variability as a result of global warming, even under low- to medium-range warming scenarios, and that these disturbances will alter the structure, composition and functioning of the system (Settele et al. 2014 <sup>[[#fn:r129|129]]</sup> ; Gauthier et al. 2015; Seddon et al. 2016 <sup>[[#fn:r130|130]]</sup> ).

In a warming climate, many species will be unable to track their climate niche as it moves, especially those in extensive flat landscapes with low dispersal capacity and in the tropics whose thermal optimum is already near current temperature (Diffenbaugh and Field 2013 <sup>[[#fn:r131|131]]</sup> ; Warszawski et al. 2013 <sup>[[#fn:r132|132]]</sup> ). Range expansion in higher latitudes and elevations as a result of warming often, but not exclusively, occurs in abandoned lands (Harsch et al. 2009 <sup>[[#fn:r133|133]]</sup> ; Landhäusser et al. 2010 <sup>[[#fn:r134|134]]</sup> ; Gottfried et al. 2012 <sup>[[#fn:r135|135]]</sup> ; Boisvert-Marsh et al. 2014 <sup>[[#fn:r136|136]]</sup> ; Bryn and Potthoff 2018 <sup>[[#fn:r137|137]]</sup> ; Rumpf et al. 2018 <sup>[[#fn:r138|138]]</sup> ; Buitenwerf et al. 2018 <sup>[[#fn:r139|139]]</sup> ; Steinbauer et al. 2018 <sup>[[#fn:r140|140]]</sup> ). This expansion typically favours thermophilic species at the expense of cold adapted species as the climate becomes suitable for lower latitude/altitude species (Rumpf et al. 2018 <sup>[[#fn:r141|141]]</sup> ). In temperate drylands, however, range expansion can be countered by intense and frequent drought conditions which result in accelerated rates of taxonomic change and spatial heterogeneity in an ecotone (Tietjen et al. 2017 <sup>[[#fn:r142|142]]</sup> ).

Since the advent of satellite observation platforms, a global increase in vegetation photosynthetic activity (i.e., greening) as evidenced through remotely sensed indices such as leaf area index (LAI) and normalised difference vegetation index (NDVI). Three satellite-based leaf area index records (GIMMS3g, GLASS and GLOMAP) imply increased growing season LAI (greening) over 25–50% and browning over less than 4% of the global vegetated area, resulting in greening trend of 0.068 ± 0.045 m <sup>2</sup> m <sup>−2</sup> yr <sup>−1</sup> over 1982–2009 (Zhu et al. 2016 <sup>[[#fn:r143|143]]</sup> ). Greening has been observed in southern Amazonia, southern Australia, the Sahel and central Africa, India, eastern China and the northern extratropical latitudes (Myneni et al. 1997 <sup>[[#fn:r144|144]]</sup> ; de Jong et al. 2012 <sup>[[#fn:r145|145]]</sup> ; Los 2013 <sup>[[#fn:r146|146]]</sup> ; Piao et al. 2015 <sup>[[#fn:r147|147]]</sup> ; Mao et al. 2016 <sup>[[#fn:r148|148]]</sup> ; Zhu et al. 2016 <sup>[[#fn:r149|149]]</sup> ; Carlson et al. 2017 <sup>[[#fn:r150|150]]</sup> ; Forzieri et al. 2017 <sup>[[#fn:r151|151]]</sup> ; Pan et al. 2018 <sup>[[#fn:r152|152]]</sup> ; Chen et al. 2019 <sup>[[#fn:r153|153]]</sup> ). Greening has been attributed to direct factors, namely human land use management and indirect factors such as CO <sub>2</sub> fertilisation, climate change, and nitrogen deposition (Donohue et al. 2013 <sup>[[#fn:r154|154]]</sup> ; Keenan et al. 2016 <sup>[[#fn:r155|155]]</sup> ; Zhu et al. 2016 <sup>[[#fn:r156|156]]</sup> ). Indirect factors have been used to explain most greening trends primarily through CO <sub>2</sub> fertilisation in the tropics and through an extended growing season and increased growing season temperatures as a result of climate change in the high latitudes (Fensholt et al. 2012 <sup>[[#fn:r157|157]]</sup> ; Zhu et al. 2016 <sup>[[#fn:r158|158]]</sup> ). The extension of the growing season in high latitudes has occurred together with an earlier spring greenup (the time at which plants begin to produce leaves in northern mid- and high-latitude ecosystems) (Goetz et al. 2015 <sup>[[#fn:r159|159]]</sup> ; Xu et al. 2016a <sup>[[#fn:r160|160]]</sup> , 2018 <sup>[[#fn:r161|161]]</sup> ) with subsequent earlier spring carbon uptake (2.3 days per decade) and gross primary productivity (GPP) (Pulliainen et al. 2017 <sup>[[#fn:r162|162]]</sup> ). The role of direct factors of greening are being increasingly investigated and a recent study has attributed over a third of observed global greening between 2000 and 2017 to direct factors, namely afforestation and croplands, in China and India (Chen et al. 2019 <sup>[[#fn:r163|163]]</sup> ).

It should be noted that measured greening is a product of satellite- derived radiance data and, as such, does not provide information on ecosystem health indicators such as species composition and richness, homeostasis, absence of disease, vigour, system resilience and the different components of ecosystems (Jørgensen et al. 2016 <sup>[[#fn:r164|164]]</sup> ). For example, a regional greening attributable to croplands expansion or intensification might occur at the expense of ecosystem biodiversity.

Within the global greening trend are also detected regional decreases in vegetation photosynthetic activity (i.e., browning) in northern Eurasia, the southwestern USA, boreal forests in North America, inner Asia and the Congo Basin, largely as a result of intensified drought stress. Since the late 1990s rates and extents of browning have exceeded those of greening in some regions, the collective result of which has been a slowdown of the global greening rate (de Jong et al. 2012 <sup>[[#fn:r165|165]]</sup> ; Pan et al. 2018 <sup>[[#fn:r166|166]]</sup> ). Within these long-term trends, inter-annual variability of regional greening and browning is attributable to regional climate variability, responses to extremes such as drought, disease and insect infestation and large- scale tele-connective controls such as ENSO and the Atlantic Multi- decadal Organization (Verbyla 2008 <sup>[[#fn:r167|167]]</sup> ; Revadekar et al. 2012 <sup>[[#fn:r168|168]]</sup> ; Epstein et al. 2018 <sup>[[#fn:r169|169]]</sup> ; Zhao et al. 2018 <sup>[[#fn:r170|170]]</sup> ).

Projected increases in drought conditions in many regions suggest long-term global vegetation greening trends are at risk of reversal to browning in a warmer climate (de Jong et al. 2012 <sup>[[#fn:r171|171]]</sup> ; Pan et al. 2018 <sup>[[#fn:r172|172]]</sup> ; Pausas and Millán 2018 <sup>[[#fn:r173|173]]</sup> ). On the other hand, in higher latitudes vegetation productivity is projected to increase as a result of higher atmospheric CO <sub>2</sub> concentrations and longer growing periods as a result of warming (Ito et al. 2016) (Section 2.3 and Box 2.3). Additionally, climate-driven transitions of ecosystems, particularly range changes, can take years to decades for the equilibrium state to be realised and the rates of these ‘committed ecosystem changes’ (Jones et al. 2009 <sup>[[#fn:r174|174]]</sup> ) vary between low and high latitudes (Jones et al. 2010 <sup>[[#fn:r175|175]]</sup> ). Furthermore, as direct factors are poorly integrated into Earth systems models (ESMs) uncertainties in projected trends of greening and browning are further compounded (Buitenwerf et al. 2018 <sup>[[#fn:r176|176]]</sup> ; Chen et al. 2019 <sup>[[#fn:r177|177]]</sup> ). Therefore, there is ''low confidence'' in the projection of global greening and browning trends.

Increased atmospheric CO <sub>2</sub> concentrations have both direct and indirect effects on terrestrial ecosystems (Sections 2.2.2 and 2.2.3, and Box 2.3). The direct effect is primarily through increased vegetation photosynthetic activity as described above. Indirect effects include decreased evapotranspiration that may offset the projected impact of drought in some water-stressed plants through improved water use efficiency in temperate regions, suggesting that some rain-fed cropping systems and grasslands will benefit from elevated atmospheric CO <sub>2</sub> concentrations (Roy et al. 2016 <sup>[[#fn:r178|178]]</sup> ; Milly and Dunne 2016 <sup>[[#fn:r179|179]]</sup> ; Swann et al. 2016 <sup>[[#fn:r180|180]]</sup> ; Chang et al. 2017 <sup>[[#fn:r181|181]]</sup> ; Zhu et al. 2017 <sup>[[#fn:r182|182]]</sup> ). In tropical regions, increased flowering activity is associated primarily with increasing atmospheric CO <sub>2</sub> , suggesting that a long- term increase in flowering activity may persist in some vegetation, particularly mid-story trees and tropical shrubs, and may enhance reproduction levels until limited by nutrient availability or climate factors such as drought frequency, rising temperatures, and reduced insolation (Pau et al. 2018 <sup>[[#fn:r183|183]]</sup> ).

<span id="climate-extremes-and-their-impact-on-land-functioning"></span>
=== 2.2.5 Climate extremes and their impact on land functioning ===

<div id="section-2-2-5-climate-extremes-and-their-impact-on-land-functioning-block-1"></div>

Extreme weather events are generally defined as the upper or lower statistical tails of the observed range of values of climate variables or climate indicators (e.g., temperature/rainfall or drought/aridity indices respectively). Previous IPCC reports have reported with ''high confidence'' on the increase of many types of observed extreme temperature events (Seneviratne et al. 2012 <sup>[[#fn:r183|183]]</sup> ; Hartmann et al. 2013b <sup>[[#fn:r184|184]]</sup> ; Hoegh-Guldberg et al. 2018 <sup>[[#fn:r185|185]]</sup> ). However, as a result of observational constraints, increases in precipitation extremes are ''less confident'' , except in observations-rich regions with dense, long-lived station networks, such as Europe and North America, where there have been likely increases in the frequency or intensity of heavy rainfall.

Extreme events occur across a wide range of time and space scales (Figure 2.3) and may include individual, relatively short-lived weather events (e.g., extreme thunderstorms storms) or a combination or accumulation of non-extreme events (Colwell et al. 2008b <sup>[[#fn:r186|186]]</sup> ; Handmer et al. 2012 <sup>[[#fn:r187|187]]</sup> ), for example, moderate rainfall in a saturated catchment having the flood peak at mean high tide (Leonard et al. 2014 <sup>[[#fn:r188|188]]</sup> ). Combinatory processes leading to a significant impact are referred to as a compound event and are a function of the nature and number of physical climate and land variables, biological agents such as pests and disease, the range of spatial and temporal scales, the strength of dependence between processes and the perspective of the stakeholder who defines the impact (Leonard et al. 2014 <sup>[[#fn:r189|189]]</sup> ; Millar and Stephenson 2015 <sup>[[#fn:r190|190]]</sup> ). Currently, there is ''low confidence'' in the impact of compound events on land as the multi-disciplinary approaches needed to address the problem are few (Zscheischler et al. 2018 <sup>[[#fn:r191|191]]</sup> ) and the rarity of compound extreme climatic events renders the analysis of impacts difficult.

<div id="section-2-2-5-climate-extremes-and-their-impact-on-land-functioning-block-2"></div>

<span id="figure-2.3"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 2.3'''

<span id="spatial-and-temporal-scales-of-typical-extreme-weather-and-climate-events-and-the-biological-systems-they-impact-shaded-grey.individuals-populations-and-ecosystems-within-these-space-time-ranges-respond-to-relevant-climate-stressors.-orange-blue-labels-indicate-an-increase-decrease-in-the-frequency-or-intensity-of-the-event-with-bold-font-reflecting-confidence-in-the-change.-non-bold"></span>
<!-- IMG CAPTION -->
'''Spatial and temporal scales of typical extreme weather and climate events and the biological systems they impact (shaded grey).Individuals, populations and ecosystems within these space-time ranges respond to relevant climate stressors. Orange (blue) labels indicate an increase (decrease) in the frequency or intensity of the event, with bold font reflecting confidence in the change. Non-bold […]'''
<!-- IMG FILE -->
[[File:1c706d64ed3ffc0e76a790971e467699 Figure-2.3-1024x612.jpg]]

Spatial and temporal scales of typical extreme weather and climate events and the biological systems they impact (shaded grey).Individuals, populations and ecosystems within these space-time ranges respond to relevant climate stressors. Orange (blue) labels indicate an increase (decrease) in the frequency or intensity of the event, with bold font reflecting confidence in the change. Non-bold black labels indicate low confidence in observed changes in frequency or intensity of these events. Each event type indicated in the figure is likely to affect biological systems at all temporal and spatial scales located to the left and below the specific event position in the figure. From Ummenhofer and Meehl (2017 <sup>[[#fn:r192|192]]</sup> ).

<!-- END IMG -->
<div id="section-2-2-5-1-changes-in-extreme-temperatures-heatwaves-and-drought"></div>

<span id="changes-in-extreme-temperatures-heatwaves-and-drought"></span>
==== 2.2.5.1 Changes in extreme temperatures, heatwaves and drought ====

<div id="section-2-2-5-1-changes-in-extreme-temperatures-heatwaves-and-drought-block-1"></div>

It is ''very likely'' that most land areas have experienced a decrease in the number of cold days and nights, and an increase in the number of warm days and unusually hot nights (Orlowsky and Seneviratne 2012 <sup>[[#fn:r193|193]]</sup> ; Seneviratne et al. 2012 <sup>[[#fn:r194|194]]</sup> ; Mishra et al. 2015 <sup>[[#fn:r195|195]]</sup> ; Ye et al. 2018 <sup>[[#fn:r196|196]]</sup> ). Although there is no consensus definition of heatwaves, as some heatwave indices have relative thresholds and others absolute thresholds, trends between indices of the same type show that recent heat-related events have been made more frequent or more intense due to anthropogenic GHG emissions in most land regions (Lewis and Karoly 2013 <sup>[[#fn:r197|197]]</sup> ; Smith et al. 2013b <sup>[[#fn:r198|198]]</sup> ; Scherer and Diffenbaugh 2014 <sup>[[#fn:r199|199]]</sup> ; Fischer and Knutti 2015 <sup>[[#fn:r200|200]]</sup> ; Ceccherini et al. 2016 <sup>[[#fn:r201|201]]</sup> ; King et al. 2016 <sup>[[#fn:r202|202]]</sup> ; Bador et al. 2016 <sup>[[#fn:r203|203]]</sup> ; Stott et al. 2016 <sup>[[#fn:r204|204]]</sup> ; King 2017 <sup>[[#fn:r205|205]]</sup> ; Hoegh-Guldberg et al. 2018 <sup>[[#fn:r206|206]]</sup> ). Globally, 50–80% of the land fraction is projected to experience significantly more intense hot extremes than historically recorded (Fischer and Knutti 2014 <sup>[[#fn:r207|207]]</sup> ; Diffenbaugh et al. 2015 <sup>[[#fn:r208|208]]</sup> ; Seneviratne et al. 2016 <sup>[[#fn:r209|209]]</sup> ). There is ''high confidence'' that heatwaves will increase in frequency, intensity and duration into the 21st century (Russo et al. 2016 <sup>[[#fn:r210|210]]</sup> ; Ceccherini et al. 2017 <sup>[[#fn:r211|211]]</sup> ; Herrera-Estrada and Sheffield 2017 <sup>[[#fn:r212|212]]</sup> ) and under high emission scenarios, heatwaves by the end of the century may become extremely long (more than 60 consecutive days) and frequent (once every two years) in Europe, North America, South America, Africa, Indonesia, the Middle East, South and Southeast Asia and Australia (Rusticucci 2012 <sup>[[#fn:r213|213]]</sup> ; Cowan et al. 2014 <sup>[[#fn:r214|214]]</sup> ; Russo et al. 2014 <sup>[[#fn:r215|215]]</sup> ; Scherer and Diffenbaugh 2014 <sup>[[#fn:r216|216]]</sup> ; Pal and Eltahir 2016 <sup>[[#fn:r217|217]]</sup> ; Rusticucci et al. 2016 <sup>[[#fn:r218|218]]</sup> ; Schär 2016 <sup>[[#fn:r219|219]]</sup> ; Teng et al. 2016 <sup>[[#fn:r220|220]]</sup> ; Dosio 2017 <sup>[[#fn:r221|221]]</sup> ; Mora et al. 2017 <sup>[[#fn:r222|222]]</sup> ; Dosio et al. 2018 <sup>[[#fn:r223|223]]</sup> ; Lehner et al. 2018 <sup>[[#fn:r224|224]]</sup> ; Lhotka et al. 2018 <sup>[[#fn:r225|225]]</sup> ; Lopez et al. 2018 <sup>[[#fn:r226|226]]</sup> ; Tabari and Willems 2018 <sup>[[#fn:r227|227]]</sup> ). Furthermore, unusual heatwave conditions today will occur regularly by 2040 under the RCP 8.5 scenario (Russo et al. 2016 <sup>[[#fn:r228|228]]</sup> ). The intensity of heat events may be modulated by land cover and soil characteristics (Miralles et al. 2014 <sup>[[#fn:r229|229]]</sup> ; Lemordant et al. 2016 <sup>[[#fn:r230|230]]</sup> ; Ramarao et al. 2016 <sup>[[#fn:r231|231]]</sup> ). Where temperature increase results in decreased soil moisture, latent heat flux is reduced while sensible heat fluxes are increased, allowing surface air temperature to rise further. However, this feedback may be diminished if the land surface is irrigated through enhanced evapotranspiration (Mueller et al. 2015 <sup>[[#fn:r232|232]]</sup> ; Siebert et al. 2017 <sup>[[#fn:r233|233]]</sup> ) (Section 2.5.2.2).

Drought (IPCC 2013c <sup>[[#fn:r234|234]]</sup> ), including megadroughts of the last century, for example, the Dustbowl drought (Hegerl et al. 2018 <sup>[[#fn:r235|235]]</sup> ) (Chapter 5), is a normal component of climate variability (Hoerling et al. 2010; Dai 2011 <sup>[[#fn:r236|236]]</sup> ) and may be seasonal, multi-year (Van Dijk et al. 2013 <sup>[[#fn:r237|237]]</sup> ) or multi-decadal (Hulme 2001 <sup>[[#fn:r238|238]]</sup> ) with increasing degrees of impact on regional activities. This inter-annual variability is controlled particularity through remote sea surface temperature (SST) forcings, such as the Inter-decadal Pacific Oscillation (IPO) and the Atlantic Multi-decadal Oscillation (AMO), El Niño/Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD), that cause drought as a result of reduced rainfall (Kelley et al. 2015 <sup>[[#fn:r239|239]]</sup> ; Dai 2011 <sup>[[#fn:r240|240]]</sup> ; Hoell et al. 2017 <sup>[[#fn:r241|241]]</sup> ; Espinoza et al. 2018 <sup>[[#fn:r242|242]]</sup> ). In some cases however, large scale SST modes do not fully explain the severity of drought some recent event attribution studies have identified a climate change fingerprint in several regional droughts, for example, the western Amazon (Erfanian et al. 2017 <sup>[[#fn:r243|243]]</sup> ), southern Africa (Funk et al. 2018 <sup>[[#fn:r244|244]]</sup> ; Yuan et al. 2018 <sup>[[#fn:r245|245]]</sup> ), southern Europe and the Mediterranean including North Africa (Kelley et al. 2015 <sup>[[#fn:r246|246]]</sup> ; Wilcox et al. 2018 <sup>[[#fn:r247|247]]</sup> ), parts of North America (Williams et al. 2015 <sup>[[#fn:r248|248]]</sup> ; Mote et al. 2016 <sup>[[#fn:r249|249]]</sup> ), Russia (Otto et al. 2012 <sup>[[#fn:r250|250]]</sup> ), India (Ramarao et al. 2015 <sup>[[#fn:r251|251]]</sup> ) and Australia (Lewis and Karoly 2013 <sup>[[#fn:r252|252]]</sup> ).

Long-term global trends in drought are difficult to determine because of this natural variability, potential deficiencies in drought indices (especially in how evapotranspiration is treated) and the quality and availability of precipitation data (Sheffield et al. 2012 <sup>[[#fn:r253|253]]</sup> ; Dai 2013 <sup>[[#fn:r254|254]]</sup> ; Trenberth et al. 2014 <sup>[[#fn:r255|255]]</sup> ; Nicholls and Seneviratne 2015 <sup>[[#fn:r256|256]]</sup> ; Mukherjee et al. 2018 <sup>[[#fn:r257|257]]</sup> ). However, regional trends in frequency and intensity of drought are evident in several parts of the world, particularly in low latitude land areas, such as the Mediterranean, North Africa and the Middle East (Vicente-Serrano et al. 2014 <sup>[[#fn:r258|258]]</sup> ; Spinoni et al. 2015a <sup>[[#fn:r259|259]]</sup> ; Dai and Zhao 2017 <sup>[[#fn:r260|260]]</sup> ; Páscoa et al. 2017 <sup>[[#fn:r261|261]]</sup> ), many regions of sub-Saharan Africa (Masih et al. 2014 <sup>[[#fn:r262|262]]</sup> ; Dai and Zhao 2017 <sup>[[#fn:r263|263]]</sup> ), central China (Wang et al. 2017e <sup>[[#fn:r264|264]]</sup> ), the southern Amazon (Fu et al. 2013 <sup>[[#fn:r265|265]]</sup> ; Espinoza et al. 2018 <sup>[[#fn:r266|266]]</sup> ), India (Ramarao et al. 2016 <sup>[[#fn:r267|267]]</sup> ), east and south Asia, parts of North America and eastern Australia (Dai and Zhao 2017 <sup>[[#fn:r268|268]]</sup> ). A recent analysis of 4500 meteorological droughts globally found increased drought frequency over the East Coast of the USA, Amazonia and north-eastern Brazil, Patagonia, the Mediterranean region, most of Africa and north-eastern China with decreased drought frequency over northern Argentina, Uruguay and northern Europe (Spinoni et al. 2019 <sup>[[#fn:r269|269]]</sup> ). The study also found drought intensity has become more severe over north-western USA, parts of Patagonia and southern Chile, the Sahel, the Congo River basin, southern Europe, north-eastern China, and south-eastern Australia, whereas the eastern USA, south-eastern Brazil, northern Europe, and central- northern Australia experienced less severe droughts. In addition to the IPCC SR15 assessment of ''medium confidence'' in increased drying over the Mediterranean region (Hoegh-Guldberg et al. 2018 <sup>[[#fn:r270|270]]</sup> ), it is further assessed with ''medium confidence'' that frequency and intensity of droughts in Amazonia, north-eastern Brazil, Patagonia, most of Africa, and north-eastern China has increased.

There is ''low confidence'' in how large-scale modes of variability will respond to a warming climate (Deser et al. 2012 <sup>[[#fn:r271|271]]</sup> ; Liu 2012 <sup>[[#fn:r272|272]]</sup> ; Christensen et al. 2013 <sup>[[#fn:r273|273]]</sup> ; Hegerl et al. 2015 <sup>[[#fn:r274|274]]</sup> ; Newman et al. 2016 <sup>[[#fn:r275|275]]</sup> ). Although, there is evidence for an increased frequency of extreme ENSO events, such as the 1997/98 El Niño and 1988/89 La Niña (Cai et al. 2014a <sup>[[#fn:r276|276]]</sup> , 2015 <sup>[[#fn:r277|277]]</sup> ) and extreme positive phases of the IOD (Christensen et al. 2013 <sup>[[#fn:r278|278]]</sup> ; Cai et al. 2014b <sup>[[#fn:r279|279]]</sup> ). However, the assessment by the SR15 was retained on an increased regional drought risk ( ''medium confidence'' ), specifically over the Mediterranean and South Africa at both 1.5°C and 2°C warming levels compared to present day, with drought risk at 2°C being significantly higher than at 1.5°C (Hoegh-Guldberg et al. 2018 <sup>[[#fn:r280|280]]</sup> ).

<div id="section-2-2-5-2-impacts-of-heat-extremes-and-drought-on-land"></div>

<span id="impacts-of-heat-extremes-and-drought-on-land"></span>
==== 2.2.5.2 Impacts of heat extremes and drought on land ====

<div id="section-2-2-5-2-impacts-of-heat-extremes-and-drought-on-land-block-1"></div>

There is ''high confidence'' that heat extremes such as unusually hot nights, extremely high daytime temperatures, heatwaves and drought are damaging to crop production (Chapter 5). Extreme heat events impact a wide variety of tree functions including reduced photosynthesis, increased photooxidative stress, leaves abscise, a decreased growth rate of remaining leaves and decreased growth of the whole tree (Teskey et al. 2015 <sup>[[#fn:r281|281]]</sup> ). Although trees are more resilient to heat stress than grasslands (Teuling et al. 2010 <sup>[[#fn:r282|282]]</sup> ), it has been observed that different types of forest (e.g., needleleaf vs broadleaf) respond differently to drought and heatwaves (Babst et al. 2012 <sup>[[#fn:r283|283]]</sup> ). For example, in the Turkish Anatolian forests net primary productivity (NPP) generally decreased during drought and heatwave events between 2000 and 2010 but in a few other regions, NPP of needle leaf forests increased (Erşahin et al. 2016 <sup>[[#fn:r284|284]]</sup> ). However, forests may become less resilient to heat stress in future due to the long recovery period required to replace lost biomass and the projected increased frequency of heat and drought events (Frank et al. 2015a <sup>[[#fn:r285|285]]</sup> ; McDowell and Allen 2015 <sup>[[#fn:r286|286]]</sup> ; Johnstone et al. 2016 <sup>[[#fn:r287|287]]</sup> ; Stevens-Rumann et al. 2018 <sup>[[#fn:r288|288]]</sup> ). Additionally, widespread regional tree mortality may be triggered directly by drought and heat stress (including warm winters) and exacerbated by insect outbreak and fire (Neuvonen et al. 1999 <sup>[[#fn:r289|289]]</sup> ; Breshears et al. 2005 <sup>[[#fn:r290|290]]</sup> ; Berg et al. 2006 <sup>[[#fn:r291|291]]</sup> ; Soja et al. 2007 <sup>[[#fn:r292|292]]</sup> ; Kurz et al. 2008 <sup>[[#fn:r293|293]]</sup> ; Allen et al. 2010 <sup>[[#fn:r294|294]]</sup> ).

Gross primary production (GPP) and soil respiration form the first and second largest carbon fluxes from terrestrial ecosystems to the atmosphere in the global carbon cycle (Beer et al. 2010 <sup>[[#fn:r295|295]]</sup> ; Bond- Lamberty and Thomson 2010 <sup>[[#fn:r296|296]]</sup> ). Heat extremes impact the carbon cycle through altering these and change ecosystem-atmosphere CO <sub>2</sub> fluxes and the ecosystem carbon balance. Compound heat and drought events result in a stronger carbon sink reduction compared to single-factor extremes as GPP is strongly reduced and ecosystem respiration less so (Reichstein et al. 2013 <sup>[[#fn:r297|297]]</sup> ; Von Buttlar et al. 2018 <sup>[[#fn:r298|298]]</sup> ). In forest biomes, however, GPP may increase temporarily as a result of increased insolation and photosynthetic activity as was seen during the 2015–2016 ENSO related drought over Amazonia (Zhu et al. 2018 <sup>[[#fn:r299|299]]</sup> ). Longer extreme events (heatwave or drought or both) result in a greater reduction in carbon sequestration and may also reverse long-term carbon sinks (Ciais et al. 2005 <sup>[[#fn:r300|300]]</sup> ; Phillips et al. 2009 <sup>[[#fn:r301|301]]</sup> ; Wolf et al. 2016b <sup>[[#fn:r302|302]]</sup> ; Ummenhofer and Meehl 2017 <sup>[[#fn:r303|303]]</sup> ; Von Buttlar et al. 2018 <sup>[[#fn:r304|304]]</sup> ; Reichstein et al. 2013 <sup>[[#fn:r305|305]]</sup> ). Furthermore, extreme heat events may impact the carbon cycle beyond the lifetime of the event. These lagged effects can slow down or accelerate the carbon cycle: it will slow down if reduced vegetation productivity and/or widespread mortality after an extreme drought are not compensated by regeneration, or speed up if productive tree and shrub seedlings cause rapid regrowth after windthrow or fire (Frank et al. 2015a <sup>[[#fn:r306|306]]</sup> ). Although some ecosystems may demonstrate resilience to a single heat climate stressor like drought (e.g., forests), compound effects of, for example, deforestation, fire and drought, potentially can result in changes to regional precipitation patterns and river discharge, losses of carbon storage and a transition to a disturbance-dominated regime (Davidson et al. 2012 <sup>[[#fn:r307|307]]</sup> ). Additionally, adaptation to seasonal drought may be overwhelmed by multi-year drought and their legacy effects (Brando et al. 2008 <sup>[[#fn:r308|308]]</sup> ; da Costa et al. 2010 <sup>[[#fn:r309|309]]</sup> ).

Under medium- and high-emission scenarios, global warming will exacerbate heat stress, thereby amplifying deficits in soil moisture and runoff despite uncertain precipitation changes (Ficklin and Novick 2017 <sup>[[#fn:r310|310]]</sup> ; Berg and Sheffield 2018 <sup>[[#fn:r311|311]]</sup> ; Cook et al. 2018 <sup>[[#fn:r312|312]]</sup> ; Dai et al. 2018 <sup>[[#fn:r313|313]]</sup> ; Engelbrecht et al. 2015 <sup>[[#fn:r314|314]]</sup> ; Ramarao et al. 2015 <sup>[[#fn:r315|315]]</sup> ; Grillakis 2019 <sup>[[#fn:r316|316]]</sup> ). This will  increase the rate of drying causing drought to set in quicker, become more intense and widespread, last longer and could result in an increased global aridity (Dai 2011 <sup>[[#fn:r317|317]]</sup> ; Prudhomme et al. 2014 <sup>[[#fn:r318|318]]</sup> ).

The projected changes in the frequency and intensity of extreme temperatures and drought is expected to result in decreased carbon sequestration by ecosystems and degradation of ecosystems health and loss of resilience (Trumbore et al. 2015 <sup>[[#fn:r319|319]]</sup> ). Also affected are many aspects of land functioning and type including agricultural productivity (Lesk et al. 2016 <sup>[[#fn:r320|320]]</sup> ), hydrology (Mosley 2015 <sup>[[#fn:r321|321]]</sup> ; Van Loon and Laaha 2015 <sup>[[#fn:r322|322]]</sup> ), vegetation productivity and distribution (Xu et al. 2011 <sup>[[#fn:r323|323]]</sup> ; Zhou et al. 2014 <sup>[[#fn:r324|324]]</sup> ), carbon fluxes and stocks, and other biogeochemical cycles (Frank et al. 2015b <sup>[[#fn:r325|325]]</sup> ; Doughty et al. 2015 <sup>[[#fn:r326|326]]</sup> ; Schlesinger et al. 2016 <sup>[[#fn:r327|327]]</sup> ). Carbon stocks are particularly vulnerable to extreme events due to their large carbon pools and fluxes, potentially large lagged impacts and long recovery times to regain lost stocks (Frank et al. 2015a <sup>[[#fn:r328|328]]</sup> ) (Section 2.2).

<div id="section-2-2-5-3-changes-in-heavy-precipitation"></div>

<span id="changes-in-heavy-precipitation"></span>
==== 2.2.5.3 Changes in heavy precipitation ====

<div id="section-2-2-5-3-changes-in-heavy-precipitation-block-1"></div>

A large number of extreme rainfall events have been documented over the past decades (Coumou and Rahmstorf 2012 <sup>[[#fn:r329|329]]</sup> ; Seneviratne et al. 2012 <sup>[[#fn:r330|330]]</sup> ; Trenberth 2012 <sup>[[#fn:r331|331]]</sup> ; Westra et al. 2013 <sup>[[#fn:r332|332]]</sup> ; Espinoza et al. 2014 <sup>[[#fn:r333|333]]</sup> ; Guhathakurta et al. 2017 <sup>[[#fn:r334|334]]</sup> ; Taylor et al. 2017 <sup>[[#fn:r335|335]]</sup> ; Thompson et al. 2017 <sup>[[#fn:r336|336]]</sup> ; Zilli et al. 2017 <sup>[[#fn:r337|337]]</sup> ). The observed shift in the trend distribution of precipitation extremes is more distinct than for annual mean precipitation and the global land fraction experiencing more intense precipitation events is larger than expected from internal variability (Fischer and Knutti 2014 <sup>[[#fn:r338|338]]</sup> ; Espinoza et al. 2018 <sup>[[#fn:r339|339]]</sup> ; Fischer et al. 2013 <sup>[[#fn:r340|340]]</sup> ). As a result of global warming, the number of record-breaking rainfall events globally has increased significantly by 12% during the period 1981–2010 compared to those expected due to natural multi-decadal climate variability (Lehmann et al. 2015 <sup>[[#fn:r341|341]]</sup> ). The IPCC SR15 reports robust increases in observed precipitation extremes for annual maximum 1-day precipitation (RX1day) and consecutive 5-day precipitation (RX5day) (Hoegh-Guldberg et al. 2018 <sup>[[#fn:r342|342]]</sup> ; Schleussner et al. 2017 <sup>[[#fn:r343|343]]</sup> ). A number of extreme rainfall events have been attributed to human influence (Min et al. 2011 <sup>[[#fn:r344|344]]</sup> ; Pall et al. 2011 <sup>[[#fn:r345|345]]</sup> ; Sippel and Otto 2014 <sup>[[#fn:r346|346]]</sup> ; Trenberth et al. 2015 <sup>[[#fn:r347|347]]</sup> ; Krishnan et al. 2016 <sup>[[#fn:r348|348]]</sup> ) and the largest fraction of anthropogenic influence is evident in the most rare and extreme events (Fischer and Knutti 2014 <sup>[[#fn:r349|349]]</sup> ).

A warming climate is expected to intensify the hydrological cycle as a warmer climate facilitates more water vapour in the atmosphere, as approximated by the Clausius-Clapeyron (C-C) relationship, with subsequent effects on regional extreme precipitation events (Christensen and Christensen 2003 <sup>[[#fn:r350|350]]</sup> ; Pall et al. 2007 <sup>[[#fn:r351|351]]</sup> ; Berg et al. 2013 <sup>[[#fn:r352|352]]</sup> ; Wu et al. 2013 <sup>[[#fn:r353|353]]</sup> ; Guhathakurta et al. 2017 <sup>[[#fn:r354|354]]</sup> ; Thompson et al. 2017 <sup>[[#fn:r355|355]]</sup> ; Taylor et al. 2017 <sup>[[#fn:r356|356]]</sup> ; Zilli et al. 2017 <sup>[[#fn:r357|357]]</sup> ; Manola et al. 2018 <sup>[[#fn:r358|358]]</sup> ). Furthermore, changes to the dynamics of the atmosphere amplify or weaken future precipitation extremes at the regional scale (O’Gorman 2015 <sup>[[#fn:r359|359]]</sup> ; Pfahl et al. 2017 <sup>[[#fn:r360|360]]</sup> ). Continued anthropogenic warming is very likely to increase the frequency and intensity of extreme rainfall in many regions of the globe (Seneviratne et al. 2012 <sup>[[#fn:r361|361]]</sup> ; Mohan and Rajeevan 2017 <sup>[[#fn:r362|362]]</sup> ; Prein et al. 2017 <sup>[[#fn:r363|363]]</sup> ; Stott et al. 2016 <sup>[[#fn:r364|364]]</sup> ) although many general circulation models (GCMs) underestimate observed increased trends in heavy precipitation suggesting a substantially stronger intensification of future heavy rainfall than the multi-model mean (Borodina et al. 2017 <sup>[[#fn:r365|365]]</sup> ; Min et al. 2011 <sup>[[#fn:r366|366]]</sup> ). Furthermore, the response of extreme convective precipitation to warming remains uncertain because GCMs and regional climate models (RCMs) are unable to explicitly simulate sub-grid scale processes such as convection, the hydrological cycle and surface fluxes and have to rely on parameterisation schemes for this (Crétat et al. 2012 <sup>[[#fn:r367|367]]</sup> ; Rossow et al. 2013 <sup>[[#fn:r368|368]]</sup> ; Wehner 2013 <sup>[[#fn:r369|369]]</sup> ; Kooperman et al. 2014 <sup>[[#fn:r370|370]]</sup> ; O’Gorman 2015 <sup>[[#fn:r371|371]]</sup> ; Larsen et al. 2016 <sup>[[#fn:r372|372]]</sup> ; Chawla et al. 2018 <sup>[[#fn:r373|373]]</sup> ; Kooperman et al. 2018 <sup>[[#fn:r374|374]]</sup> ; Maher et al. 2018 <sup>[[#fn:r375|375]]</sup> ; Rowell and Chadwick 2018 <sup>[[#fn:r376|376]]</sup> ). High-resolution RCMs that explicitly resolve convection have a better representation of extreme precipitation but are dependent on the GCM to capture the large scale environment in which the extreme event may occur (Ban et al. 2015 <sup>[[#fn:r377|377]]</sup> ; Prein et al. 2015 <sup>[[#fn:r378|378]]</sup> ; Kendon et al. 2017 <sup>[[#fn:r379|379]]</sup> ). Inter- annual variability of precipitation extremes in the convective tropics are not well captured by global models (Allan and Liu 2018 <sup>[[#fn:r380|380]]</sup> ).

There is ''low confidence'' in the detection of long-term observed and projected seasonal and daily trends of extreme snowfall. The narrow rain–snow transition temperature range at which extreme snowfall can occur is relatively insensitive to climate warming and subsequent large interdecadal variability (Kunkel et al. 2013 <sup>[[#fn:r381|381]]</sup> ; O’Gorman 2014 <sup>[[#fn:r382|382]]</sup> , 2015 <sup>[[#fn:r383|383]]</sup> ).

<div id="section-2-2-5-4-impacts-of-precipitation-extremes-on-different-land-cover-types"></div>

<span id="impacts-of-precipitation-extremes-on-different-land-cover-types"></span>
==== 2.2.5.4 Impacts of precipitation extremes on different land cover types ====

<div id="section-2-2-5-4-impacts-of-precipitation-extremes-on-different-land-cover-types-block-1"></div>

More intense rainfall leads to water redistribution between surface and ground water in catchments as water storage in the soil decreases (green water) and runoff and reservoir inflow increases (blue water) (Liu and Yang 2010 <sup>[[#fn:r384|384]]</sup> ; Eekhout et al. 2018 <sup>[[#fn:r385|385]]</sup> ). This results in increased surface flooding and soil erosion, increased plant water stress and reduced water security, which in terms of agriculture means an increased dependency on irrigation and reservoir storage (Nainggolan et al. 2012 <sup>[[#fn:r386|386]]</sup> ; Favis-Mortlock and Mullen 2011 <sup>[[#fn:r387|387]]</sup> ; García- Ruiz et al. 2011 <sup>[[#fn:r388|388]]</sup> ; Li and Fang 2016 <sup>[[#fn:r389|389]]</sup> ; Chagas and Chaffe 2018 <sup>[[#fn:r390|390]]</sup> ). As there is high confidence of a positive correlation between global warming and future flood risk, land cover and processes are likely to be negatively impacted, particularly near rivers and in floodplains (Kundzewicz et al. 2014 <sup>[[#fn:r391|391]]</sup> ; Alfieri et al. 2016 <sup>[[#fn:r392|392]]</sup> ; Winsemius et al. 2016 <sup>[[#fn:r393|393]]</sup> ; Arnell and Gosling 2016 <sup>[[#fn:r394|394]]</sup> ; Alfieri et al. 2017 <sup>[[#fn:r395|395]]</sup> ; Wobus et al. 2017 <sup>[[#fn:r396|396]]</sup> ).

In agricultural systems, heavy precipitation and inundation can delay planting, increase soil compaction and cause crop losses through anoxia and root diseases (Posthumus et al. 2009 <sup>[[#fn:r397|397]]</sup> ). In tropical regions, flooding associated with tropical cyclones can lead to crop failure from both rainfall and storm surge. In some cases, flooding can affect yield more than drought, particularly in tropical regions (e.g., India) and in some mid/high latitude regions such as China and central and northern Europe (Zampieri et al. 2017 <sup>[[#fn:r398|398]]</sup> ). Waterlogging of croplands and soil erosion also negatively affect farm operations and block important transport routes (Vogel and Meyer 2018 <sup>[[#fn:r399|399]]</sup> ; Kundzewicz and Germany 2012 <sup>[[#fn:r400|400]]</sup> ). Flooding can be beneficial in drylands if the floodwaters infiltrate and recharge alluvial aquifers along ephemeral river pathways, extending water availability into dry seasons and drought years, and supporting riparian systems and human communities (Kundzewicz and Germany 2012; Guan et al. 2015 <sup>[[#fn:r401|401]]</sup> ). Globally, the impact of rainfall extremes on agriculture is less than that of temperature extremes and drought, although in some regions and for some crops, extreme precipitation explains a greater component of yield variability, for example, of maize in the Midwestern USA and southern Africa (Ray et al. 2015 <sup>[[#fn:r402|402]]</sup> ; Lesk et al. 2016 <sup>[[#fn:r403|403]]</sup> ; Vogel et al. 2019 <sup>[[#fn:r404|404]]</sup> ).

Although many soils on floodplains regularly suffer from inundation, the increases in the magnitude of flood events mean that new areas with no recent history of flooding are now becoming severely affected (Yellen et al. 2014 <sup>[[#fn:r405|405]]</sup> ). Surface flooding and associated soil saturation often results in decreased soil quality through nutrient loss, reduced plant productivity, stimulated microbial growth and microbial community composition, negatively impacted soil redox and increased GHG emissions (Bossio and Scow 1998 <sup>[[#fn:r406|406]]</sup> ; Niu et al. 2014 <sup>[[#fn:r407|407]]</sup> ; Barnes et al. 2018 <sup>[[#fn:r408|408]]</sup> ; Sánchez-Rodríguez et al. 2019 <sup>[[#fn:r409|409]]</sup> ). The impact of flooding on soil quality is influenced by management systems that may mitigate or exacerbate the impact. Although soils tend to recover quickly after floodwater removal, the impact of repeated extreme flood events over longer timescales on soil quality and function is unclear (Sánchez-Rodríguez et al. 2017 <sup>[[#fn:r410|410]]</sup> ).

Flooding in ecosystems may be detrimental through erosion or permanent habitat loss, or beneficial, as a flood pulse brings nutrients to downstream regions (Kundzewicz et al. 2014 <sup>[[#fn:r411|411]]</sup> ). Riparian forests can be damaged through flooding; however, increased flooding may also be of benefit to forests where upstream water demand has lowered stream flow, but this is difficult to assess and the effect of flooding on forests is not well studied (Kramer et al. 2008 <sup>[[#fn:r412|412]]</sup> ; Pawson et al. 2013 <sup>[[#fn:r413|413]]</sup> ). Forests may mitigate flooding, however flood mitigation potential is limited by soil saturation and rainfall intensity (Pilaš et al. 2011 <sup>[[#fn:r414|414]]</sup> ; Ellison et al. 2017 <sup>[[#fn:r415|415]]</sup> ). Some grassland species under heavy rainfall and soil saturated conditions responded negatively with decreased reproductive biomass and germination rates (Gellesch et al. 2017 <sup>[[#fn:r416|416]]</sup> ), however overall productivity in grasslands remains constant in response to heavy rainfall (Grant et al. 2014 <sup>[[#fn:r417|417]]</sup> ).

Extreme rainfall alters responses of soil CO <sub>2</sub> fluxes and CO <sub>2</sub> uptake by plants within ecosystems, and therefore result in changes in ecosystem carbon cycling (Fay et al. 2008 <sup>[[#fn:r418|418]]</sup> ; Frank et al. 2015a <sup>[[#fn:r419|419]]</sup> ). Extreme rainfall and flooding limits oxygen in soil which may suppress the activities of soil microbes and plant roots and lower soil respiration, therefore lowering carbon cycling (Knapp et al. 2008 <sup>[[#fn:r420|420]]</sup> ; Rich and Watt 2013 <sup>[[#fn:r421|421]]</sup> ; Philben et al. 2015 <sup>[[#fn:r422|422]]</sup> ). However, the impact of extreme rainfall on carbon fluxes in different biomes differs. For example, extreme rainfall in mesic biomes reduces soil CO <sub>2</sub> flux to the atmosphere and GPP whereas in xeric biomes the opposite is true, largely as a result of increased soil water availability (Knapp and Smith 2001 <sup>[[#fn:r423|423]]</sup> ; Heisler and Knapp 2008 <sup>[[#fn:r424|424]]</sup> ; Heisler-White et al. 2009 <sup>[[#fn:r425|425]]</sup> ; Zeppel et al. 2014 <sup>[[#fn:r426|426]]</sup> ; Xu and Wang 2016 <sup>[[#fn:r427|427]]</sup> ; Liu et al. 2017b <sup>[[#fn:r428|428]]</sup> ; Connor and Hawkes 2018 <sup>[[#fn:r429|429]]</sup> ).

As shown above GHG fluxes between the land and atmosphere are affected by climate. The next section assesses these fluxes in greater detail and the potential for land as a carbon sink.

<div id="section-2-2-5-4-impacts-of-precipitation-extremes-on-different-land-cover-types-block-2" class="box"></div>

<span id="ccb3-fire-and-climate-change"></span>