Editing IPCC:AR6/SRCCL/Chapter-4 (section)

== 4.6 Impacts of land degradation on climate ==

<div id="article-4-6-impacts-of-land-degradation-on-climate-block-1"></div>

While Chapter 2 has its focus on land cover changes and their impacts on the climate system, this chapter focuses on the influences of individual land degradation processes on climate (see Table 4.1) which may or may not take place in association with land cover changes. The effects of land degradation on CO <sub>2</sub> and other GHGs as well as those on surface albedo and other physical controls of the global radiative balance are discussed.

<span id="impact-on-greenhouse-gases-ghgs"></span>
=== 4.6.1 Impact on greenhouse gases (GHGs) ===

<div id="section-4-6-1-impact-on-greenhouse-gases-ghgs-block-1"></div>

Land degradation processes with direct impact on soil and terrestrial biota have great relevance in terms of CO <sub>2</sub> exchange with the atmosphere, given the magnitude and activity of these reservoirs in the global carbon cycle. As the most widespread form of soil degradation, erosion detaches the surface soil material, which typically hosts the highest organic carbon stocks, favouring the mineralisation and release as CO <sub>2</sub> . Yet complementary processes such as carbon burial may compensate for this effect, making soil erosion a long-term carbon sink ( ''low agreement, limited evidence'' ), (Wang et al. (2017b) <sup>[[#fn:r783|783]]</sup> , but see also Chappell et al. (2016) <sup>[[#fn:r784|784]]</sup> ). Precise estimation of the CO <sub>2</sub> released from eroded lands is challenged by the fact that only a fraction of the detached carbon is eventually lost to the atmosphere. It is important to acknowledge that a substantial fraction of the eroded material may preserve its organic carbon load in field conditions. Moreover, carbon sequestration may be favoured through the burial of both the deposited material and the surface of its hosting soil at the deposition location (Quinton et al. 2010 <sup>[[#fn:r785|785]]</sup> ). The cascading effects of erosion on other environmental processes at the affected sites can often cause net CO <sub>2</sub> emissions through their indirect influence on soil fertility, and the balance of organic carbon inputs and outputs, interacting with other non-erosive soil degradation processes (such as nutrient depletion, compaction and salinisation), which can lead to the same net carbon effects (see Table 4.1) (van de Koppel et al. 1997 <sup>[[#fn:r786|786]]</sup> ).

As natural and human-induced erosion can result in net carbon storage in very stable buried pools at the deposition locations, degradation in those locations has a high C-release potential. Coastal ecosystems such as mangrove forests, marshes and seagrasses are at typical deposition locations, and their degradation or replacement with other vegetation is resulting in a substantial carbon release (0.15 to 1.02 GtC yr <sup>–1</sup> ) (Pendleton et al. 2012 <sup>[[#fn:r787|787]]</sup> ), which highlights the need for a spatially integrated assessment of land degradation impacts on climate that considers in-situ but also ex-situ emissions.

Cultivation and agricultural management of cultivated land are relevant in terms of global CO <sub>2</sub> land–atmosphere exchange (Section 4.8.1). Besides the initial pulse of CO <sub>2</sub> emissions associated with the onset of cultivation and associated vegetation clearing (Chapter 2), agricultural management practices can increase or reduce carbon losses to the atmosphere. Although global croplands are considered to be at a relatively neutral stage in the current decade (Houghton et al. 2012 <sup>[[#fn:r788|788]]</sup> ), this results from a highly uncertain balance between coexisting net losses and gains. Degradation losses of soil and biomass carbon appear to be compensated by gains from soil protection and restoration practices such as cover crops, conservation tillage and nutrient replenishment favouring organic matter build-up. Cover crops, increasingly used to improve soils, have the potential to sequester 0.12 GtC yr <sup>–1</sup> on global croplands with a saturation time of more than 150 years (Poeplau and Don 2015 <sup>[[#fn:r789|789]]</sup> ). No-till practices (i.e., tillage elimination favouring crop residue retention in the soil surface) which were implemented to protect soils from erosion and reduce land preparation times, were also seen with optimism as a carbon sequestration option, which today is considered more modest globally and, in some systems, even less certain (VandenBygaart 2016 <sup>[[#fn:r799|799]]</sup> ; Cheesman et al. 2016 <sup>[[#fn:r791|791]]</sup> ; Powlson et al. 2014 <sup>[[#fn:r792|792]]</sup> ). Among soil fertility restoration practices, lime application for acidity correction, increasingly important in tropical regions, can generate a significant net CO <sub>2</sub> source in some soils (Bernoux et al. 2003 <sup>[[#fn:r793|793]]</sup> ; Desalegn et al. 2017 <sup>[[#fn:r794|794]]</sup> ).

Land degradation processes in seminatural ecosystems driven by unsustainable uses of their vegetation through logging or grazing lead to reduced plant cover and biomass stocks, causing net carbon releases from soils and plant stocks. Degradation by logging activities is particularly prevalent in developing tropical and subtropical regions, involving carbon releases that exceed by far the biomass of harvested products, including additional vegetation and soil sources that are estimated to reach 0.6 GtC yr <sup>–1</sup> (Pearson et al. 2014, 2017 <sup>[[#fn:r795|795]]</sup> ). Excessive grazing pressures pose a more complex picture with variable magnitudes and even signs of carbon exchanges. A general trend of higher carbon losses in humid overgrazed rangelands suggests a high potential for carbon sequestration following the rehabilitation of those systems (Conant and Paustian 2002 <sup>[[#fn:r796|796]]</sup> ) with a global potential sequestration of 0.045 GtC yr <sup>-1</sup> . A special case of degradation in rangelands is the process leading to the woody encroachment of grass-dominated systems, which can be responsible for declining animal production but high carbon sequestration rates (Asner et al. 2003 <sup>[[#fn:r797|797]]</sup> ; Maestre et al. 2009 <sup>[[#fn:r798|798]]</sup> ).

Fire regime shifts in wild and seminatural ecosystems can become a degradation process in itself, with high impact on net carbon emission and with underlying interactive human and natural drivers such as burning policies (Van Wilgen et al. 2004 <sup>[[#fn:r1651|1651]]</sup> ), biological invasions (Brooks et al. 2009 <sup>[[#fn:r800|800]]</sup> ), and plant pest/disease spread (Kulakowski et al. 2003 <sup>[[#fn:r801|801]]</sup> ). Some of these interactive processes affecting unmanaged forests have resulted in massive carbon release, highlighting how degradation feedbacks on climate are not restricted to intensively used land but can affect wild ecosystems as well (Kurz et al. 2008 <sup>[[#fn:r802|802]]</sup> ).

Agricultural land and wetlands represent the dominant source of non-CO <sub>2</sub> greenhouse gases (GHGs) (Chen et al. 2018d <sup>[[#fn:r803|803]]</sup> ). In agricultural land, the expansion of rice cultivation (increasing CH <sub>4</sub> sources), ruminant stocks and manure disposal (increasing CH <sub>4</sub> , N <sub>2</sub> O and NH <sub>3</sub> fluxes) and nitrogen over-fertilisation combined with soil acidification (increasing N <sub>2</sub> O fluxes) are introducing the major impacts ( ''medium agreement, medium evidence'' ) and their associated emissions appear to be exacerbated by global warming ( ''medium agreement, medium evidence'' ) (Oertel et al. 2016 <sup>[[#fn:r804|804]]</sup> ).

As the major sources of global N <sub>2</sub> O emissions, over-fertilisation and manure disposal are not only increasing in-situ sources but also stimulating those along the pathway of dissolved inorganic nitrogen transport all the way from draining waters to the ocean ( ''high agreement, medium evidence'' ). Current budgets of anthropogenically fixed nitrogen on the Earth System (Tian et al. 2015 <sup>[[#fn:r805|805]]</sup> ; Schaefer et al. 2016 <sup>[[#fn:r806|806]]</sup> ; Wang et al. 2017a <sup>[[#fn:r807|807]]</sup> ) suggest that N <sub>2</sub> O release from terrestrial soils and wetlands accounts for 10–15% of the emissions, yet many further release fluxes along the hydrological pathway remain uncertain, with emissions from oceanic ‘dead-zones’ being a major aspect of concern (Schlesinger 2009; Rabalais et al. 2014 <sup>[[#fn:r808|808]]</sup> ).

Environmental degradation processes focused on the hydrological system, which are typically manifested at the landscape scale, include both drying (as in drained wetlands or lowlands) and wetting trends (as in waterlogged and flooded plains). Drying of wetlands reduces CH <sub>4</sub> emissions (Turetsky et al. 2014 <sup>[[#fn:r812|812]]</sup> ) but favours pulses of organic matter mineralisation linked to high N <sub>2</sub> O release (Morse and Bernhardt 2013 <sup>[[#fn:r813|813]]</sup> ; Norton et al. 2011 <sup>[[#fn:r814|814]]</sup> ). The net warming balance of these two effects is not resolved and may be strongly variable across different types of wetlands. In the case of flooding of non-wetland soils, a suppression of CO <sub>2</sub> release is typically overcompensated in terms of net greenhouse impact by enhanced CH <sub>4</sub> fluxes that stem from the lack of aeration but are aided by the direct effect of extreme wetting on the solubilisation and transport of organic substrates (McNicol and Silver 2014 <sup>[[#fn:r815|815]]</sup> ). Both wetlands rewetting/restoration and artificial wetland creation can increase CH <sub>4</sub> release (Altor and Mitsch 2006 <sup>[[#fn:r816|816]]</sup> ; Fenner et al. 2011 <sup>[[#fn:r817|817]]</sup> ). Permafrost thawing is another major source of CH <sub>4</sub> release, with substantial long-term contributions to the atmosphere that are starting to be globally quantified (Christensen et al. 2004 <sup>[[#fn:r818|818]]</sup> ; Schuur et al. 2015 <sup>[[#fn:r819|819]]</sup> ; Walter Anthony et al. 2016 <sup>[[#fn:r820|820]]</sup> ).

<span id="physical-impacts"></span>
=== 4.6.2 Physical impacts ===

<div id="section-4-6-2-physical-impacts-block-1"></div>

Among the physical effects of land degradation, surface albedo changes are those with the most evident impact on the net global radiative balance and net climate warming/cooling. Degradation processes affecting wild and semi-natural ecosystems, such as fire regime changes, woody encroachment, logging and overgrazing, can trigger strong albedo changes before significant biogeochemical shifts take place. In most cases these two types of effects have opposite signs in terms of net radiative forcing, making their joint assessment critical for understanding climate feedbacks (Bright et al. 2015 <sup>[[#fn:r821|821]]</sup> ).

In the case of forest degradation or deforestation, the albedo impacts are highly dependent on the latitudinal/climatic belt to which they belong. In boreal forests, the removal or degradation of the tree cover increases albedo (net cooling effect) ( ''medium evidence, high agreement'' ) as the reflective snow cover becomes exposed, which can exceed the net radiative effect of the associated carbon release to the atmosphere (Davin et al. 2010 <sup>[[#fn:r822|822]]</sup> ; Pinty et al. 2011 <sup>[[#fn:r823|823]]</sup> ). On the other hand, progressive greening of boreal and temperate forests has contributed to net albedo declines ( ''medium agreement, medium evidence'' ) (Planque et al. 2017 <sup>[[#fn:r824|824]]</sup> ; Li et al. 2018a <sup>[[#fn:r825|825]]</sup> ). In the northern treeless vegetation belt (tundra), shrub encroachment leads to the opposite effect as the emergence of plant structures above the snow cover level reduce winter-time albedo (Sturm 2005 <sup>[[#fn:r826|826]]</sup> ).

The extent to which albedo shifts can compensate for carbon storage shifts at the global level has not been estimated. A significant but partial compensation takes place in temperate and subtropical dry ecosystems in which radiation levels are higher and carbon stocks smaller compared to their more humid counterparts ( ''medium agreement, medium evidence'' ). In cleared dry woodlands, half of the net global warming effect of net carbon release has been compensated by albedo increase (Houspanossian et al. 2013 <sup>[[#fn:r827|827]]</sup> ), whereas in afforested dry rangelands, albedo declines cancelled one-fifth of the net carbon sequestration (Rotenberg and Yakir 2010 <sup>[[#fn:r828|828]]</sup> ). Other important cases in which albedo effects impose a partial compensation of carbon exchanges are the vegetation shifts associated with wildfires, as shown for the savannahs, shrublands and grasslands of Sub-Saharan Africa (Dintwe et al. 2017 <sup>[[#fn:r829|829]]</sup> ). Besides the net global effects discussed above, albedo shifts can play a significant role in local climate ( ''high agreement, medium evidence'' ), as exemplified by the effect of no-till agriculture reducing local heat extremes in European landscapes (Davin et al. 2014 <sup>[[#fn:r830|830]]</sup> ) and the effects of woody encroachment causing precipitation rises in the North American Great Plains (Ge and Zou 2013 <sup>[[#fn:r831|831]]</sup> ). Modelling efforts that integrate ground data from deforested areas worldwide accounting for both physical and biogeochemical effects, indicate that massive global deforestation would have a net warming impact (Lawrence and Vandecar 2015 <sup>[[#fn:r832|832]]</sup> ) at both local and global levels with highlight non-linear effects of forest loss on climate variables.

Beyond the albedo effects presented above, other physical impacts of land degradation on the atmosphere can contribute to global and regional climate change. Of particular relevance, globally and continentally, are the net cooling effects of dust emissions ( ''low agreement, medium evidence'' ) (Lau and Kim (2007) <sup>[[#fn:r833|833]]</sup> , but see also Huang et al. (2014) <sup>[[#fn:r834|834]]</sup> ). Anthropogenic emission of mineral particles from degrading land appear to have a similar radiative impact than all other anthropogenic aerosols (Sokolik and Toon 1996 <sup>[[#fn:r835|835]]</sup> ). Dust emissions may explain regional climate anomalies through reinforcing feedbacks, as suggested for the amplification of the intensity, extent and duration of the low precipitation anomaly of the North American Dust Bowl in the 1930s (Cook et al. 2009 <sup>[[#fn:r836|836]]</sup> ). Another source of physical effects on climate are surface roughness changes which, by affecting atmospheric drag, can alter cloud formation and precipitation (low agreement, low evidence), as suggested by modelling studies showing how the massive deployment of solar panels in the Sahara could increase rainfall in the Sahel (Li et al. 2018c <sup>[[#fn:r837|837]]</sup> ), or how woody encroachment in the Arctic tundra could reduce cloudiness and raise temperature (Cho et al. 2018 <sup>[[#fn:r838|838]]</sup> ). The complex physical effects of deforestation, as explored through modelling, converge into general net regional precipitation declines, tropical temperature increases and boreal temperature declines, while net global effects are less certain (Perugini et al. 2017 <sup>[[#fn:r839|839]]</sup> ). Integrating all the physical effects of land degradation and its recovery or reversal is still a challenge, yet modelling attempts suggest that, over the last three decades, the slow but persistent net global greening caused by the average increase of leaf area in the land has caused a net cooling of the Earth, mainly through the rise in evapotranspiration (Zeng et al. 2017 <sup>[[#fn:r840|840]]</sup> ) ( ''low confidence'' ).

<span id="impacts-of-climate-related-land-degradation-on-poverty-and-livelihoods"></span>