See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/344351223

A research agenda for the restoration of tropical and subtropical grasslands

and savannas

Article  in  Restoration Ecology · September 2020

DOI: 10.1111/rec.13292

CITATIONS READS

7 406

17 authors, including:

Elise Buisson Alessandra Fidelis

Avignon Université São Paulo State University
170 PUBLICATIONS   3,521 CITATIONS    112 PUBLICATIONS   4,470 CITATIONS   

SEE PROFILE SEE PROFILE

Gerhard E. Overbeck Isabel Schmidt

Universidade Federal do Rio Grande do Sul University of Brasília
173 PUBLICATIONS   5,874 CITATIONS    52 PUBLICATIONS   1,162 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Will climate change make sourveld more sour? View project

Ecological engineering to fight against soil erosion View project

All content following this page was uploaded by Truman P Young on 14 August 2021.

The user has requested enhancement of the downloaded file.

REVIEW ARTICLE

A research agenda for the restoration of tropical and

subtropical grasslands and savannas
Elise Buisson1,2,3 , Alessandra Fidelis4, Gerhard E. Overbeck5 , Isabel B. Schmidt6 ,
Giselda Durigan7 , Truman P. Young2, Swanni T. Alvarado8, André J. Arruda1,9, Sylvain Boisson10,
William Bond11, André Coutinho12, Kevin Kirkman13, Rafael S. Oliveira14, Melissa H. Schmitt15,16,
Frances Siebert17, Stefan J. Siebert17, Dave I. Thompson15,18, Fernando A. O. Silveira9

Despite growing recognition of the conservation value of grassy biomes, our understanding of how to restore biodiverse tropical
and subtropical grassy biomes (grasslands and savannas; TGB) remains limited. Several tools have recently been identified for
TGB restoration, including prescribed fires, appropriate management of livestock and wild herbivores, tree cutting and shrub
removal, invasive species control, and the reintroduction of native grasses and forbs via seeding or transplants. However, addi-
tional research for improved TGB restoration is needed. This article aims to identify ecological research priorities for TGB res-
toration. The following points are crucial to scale up TGB restoration and meet the challenges of the UN Restoration Decade.
Research should focus on: disentangling the reasons why TGB are often undervalued and misunderstood; mapping TGB res-
toration opportunities; identifying regions where TGB and other biomes naturally exist as alternative stable states; recognizing
areas with natural regeneration potential to avoid unnecessary intervention; restoring soil conditions; disentangling factors
driving low seed quality, determining germination requirements and developing vegetative propagation techniques for TGB
species; disentangling the limiting factors and key ecological processes underlying seedling establishment and community
assembly; improving and validating long-term management to mimic natural disturbance regimes; setting the minimum attri-
butes of desirable TGB in terms of structure, composition, functioning, and resilience; and improving monitoring of restoration
outcomes. Such research has the potential to advance theory, policy, and practice in TGB restoration, ultimately resulting in
long-term benefits for people and nature in some of the more neglected ecosystems of our planet.
Key words: grasslands, grassy biomes, old-growth grasslands, savannas, tropical old-growth grassy biomes, woodlands

Introduction
Implications for Practice
International incentives in the last decade have brought ecolog-
• The restoration of tropical grassy biomes (TGB) cannot
ical restoration to the forefront of proposed solutions to mitigate
be upscaled if these ecosystems remain poorly known
the impact of ecosystem degradation worldwide (Temperton
and undervalued, if their ecosystem services are not per-
et al. 2019). Ten years ago, the UN Convention on Biological
ceived or realized, and if the representation of restoration
Diversity proposed that 15% of degraded ecosystems be
opportunities remains misleading.
restored by 2020 (CBD 2011), whereas the Bonn Challenge
• Low seed quality, germination, and plant establishment
aimed specifically at “restoring” (i.e. reforesting) 150 million
must be investigated and overcome using innovative
technologies to enhance restoration practice. hectares of the world’s deforested and degraded land by 2020,
• Plant propagation techniques must be developed to facil- and 350 million hectares by 2030 (IUCN-DC 2011). These ini-
tiatives were subsequently endorsed by the New York Declara-
itate the reintroduction of the large proportion of TGB
species whose populations do not readily regenerate from tion on Forests at the 2014 UN Climate Summit (Climate
the seed bank or seed rain. Summit 2014) and by the UN Sustainable Development Goals
• Critical research also includes finding the best practices (IRP 2019). Recently, the UN Environment Program declared
2021–2030 to be the “UN Decade on Ecosystem Restoration”
for site preparation to preserve/recover soil properties.
• Monitoring should be evidence-based to assess whether (UN Environment Programme 2019). Unfortunately, such poli-
cies are strongly forest- and tree-oriented, overestimate the
desirable structure, composition, functioning, and espe-
cially resilience of a TGB have been recovered. potential of tree planting as a means of climate mitigation
(Veldman et al. 2019), and do not fully acknowledge limits

Author contributions: EB, FAOS, RSO conceived and organized the original symposium at the 2019 SER World Conference; EB, AF, GEO, IBS, GD, TPY, STA, FAOS prepared the
first draft; all authors actively discussed ideas, contributed with writing topics, revised, and approved the final version of the manuscript.

April 2021 Restoration Ecology Vol. 29, No. S1, e13292 1 of 18

Research agenda for tropical grassland restoration

linked to forest stability and important ecological and climate- broad themes in restoration to structure questions aimed at opti-
related risks (Anderegg et al. 2020). The intense pressure to mizing practical restoration outcomes (Miller et al. 2017; Fig. 1).
afforest has even been identified as a major threat to non-forest Our focus is on the ecological dimensions of restoration ecol-
ecosystems and human livelihoods (Fagan 2020; Johansson ogy, but we also recognize the need for similar efforts toward
et al. 2020; Tölgyesi et al. 2020), and provides a compelling rea- socioecological dimensions of restoration (sensu Perring
son for establishing a detailed research agenda for the restoration et al. 2015). Most of the questions we identify are from the per-
of overlooked and misinterpreted open ecosystems (Temperton spective of the vegetation, but we also address some priority
et al. 2019), including tropical grassy biomes. issues concerning herbivores in tropical grassy biomes.
Upscaling restoration, however, is challenging as it requires
knowledge of natural ecosystems and their socioeconomic con-
text, and solid links between science and practice (Miller Changing Perceptions of Tropical Grassy Biomes
et al. 2017; Cooke et al. 2019; Gann et al. 2019). Major hurdles Misunderstandings and misconceptions concerning the ecology
limiting appropriate restoration and management of old-growth of open ecosystems (Bond 2019; Fagan 2020) have led to pris-
tropical and subtropical grasslands and savannas (TGB, i.e. trop- tine TGB being confused with degraded or secondary grasslands
ical grassy biomes) have been recently identified (Buisson (e.g. TGB transformed to pastures) or with derived grasslands
et al. 2019), yet it is unclear how to achieve the deeper scientific and savannas (e.g. grasslands derived from the degradation of
knowledge necessary to overcome such hurdles. TGB are mega- forests), with subsequent recommendations for restoration
diverse, ancient ecosystems with a continuous herbaceous cover directed at tree planting (Silveira et al. 2020; Schmidt
composed of graminoids and forbs, with or without scattered et al. 2019a; Fig. 2). TGB are often thought to be by-products
shrubs and trees (Veldman et al. 2015). TGB are largely threat- of forest degradation, despite clear criteria for identifying TGB
ened by agricultural conversion, opencast mining, anthropo- being available (Veldman et al. 2015). Such misidentification
genic increases in or suppression of fires, mismanaged of TGB results in ill-conceived restoration policies, including
livestock, wood extraction, invasive species, and afforestation afforestation (Veldman 2016; Bond 2019). TGB have been
(Veldman et al. 2015). Such threats compromise biodiversity, demonstrated to be slow to recover if degraded by exogenous
ecosystem functioning, and human livelihoods. Owing to large disturbances (Nerlekar & Veldman 2020).
knowledge gaps and widespread misconceptions concerning Disentangling the reasons why TGB are often undervalued
their basic ecology (Bond 2019; Silveira et al. 2020), large-scale and misunderstood (Parr et al. 2014; Veldman et al. 2017) is a
restoration in these key ecosystems is not yet feasible, despite major research question which cannot be tackled by restoration
their rich biodiversity, provision of ecosystem services, and ecologists alone, but calls for integration with environmental
direct support of human livelihoods. Clearly, international sociology. This research is likely to provide conditions to
agreements should recognize and endorse the fact that forests change the historical prejudice against open ecosystems
are not the only ecosystems in need of restoration, to make fund- (Bond 2019) that will promote best restoration practices and pol-
ing and government incentives available for the restoration of icy. Such a changed perspective can also be supported by better
open ecosystems. Additionally, limited knowledge of the ecol- communication of the high benefit–cost ratios of TGB restora-
ogy and restoration of TGB reinforces the need for prioritization tion (De Groot et al. 2013) and by the plethora of ecosystem ser-
of research efforts to support and foster practices aimed at accel- vices provided by restored ecosystems (Morecroft et al. 2019).
erating TGB restoration. Restoration benefits and costs should be identified regionally
Here we propose a research agenda to guide research priori- to provide support for practical decisions in TGB restoration.
ties for TGB restoration. We identify emerging issues across Socioecological research is needed to demonstrate the value

1
Avignon Université, Institut Méditerranéen de Biodiversité et d’Ecologie - IMBE, CNRS, IRD, Aix Marseille Université, IUT d’Avignon, AGROPARC BP61207, Avignon cedex 9,
84911, France
2
Department of Plant Sciences, University of California, Davis, CA 95616, U.S.A.
3
Address correspondence to E. Buisson, email [email protected]
4
Lab of Vegetation Ecology, Instituto de Biociências, Universidade Estadual Paulista (UNESP), Av. 24A, 1515, Rio Claro, SP 13506-900, Brazil
5
Departamento de Botânica, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves 9500, CEP, Porto Alegre, RS 91501-970, Brazil
6
Department of Ecology, University of Brasília, Campus Universitário Darcy Ribeiro, Brasilia, Brazil
7
Floresta Estadual de Assis, Instituto Florestal do Estado de S~ao Paulo, P.O. Box 104, Assis, SP 19802-970, Brazil
8
Universidade Estadual do Maranh~ao, S~ao Luís, Maranh~ao, Brazil
9
Departamento de Genética, Ecologia e Evoluç~ao, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais, Av. Antônio Carlos 6627, CEP, Belo Horizonte, MG 31270-
901, Brazil
10
Université de Liège Gembloux Agro-Bio Tech, Biodiversity and Landscape, TERRA, Gembloux, Belgium
11
Department of Biological Sciences, University of Cape Town, Rondebosch, South Africa
12
Ecology Graduate Program, University of Brasília, Campus Universitário Darcy Ribeiro, Brasília, DF 70.910-900, Brazil
13
School of Life Science, University of KwaZulu-Natal, Pietermaritzburg, KwaZulu-Natal, South Africa
14
Department of Plant Biology, Institute of Biology, University of Campinas – UNICAMP, Campinas, SP, Brazil
15
South African Environmental Observation Network, Ndlovu Node, Scientific Services, Kruger National Park, Private Bag X1021, Phalaborwa, 1390, South Africa
16
Department of Ecology, Evolution, and Marine Biology, University of California Santa Barbara, Santa Barbara, CA, U.S.A.
17
Unit for Environmental Sciences and Management, North-West University, 11 Hoffman Street, Potchefstroom, North-West 2531, South Africa
18
School of Geography, Archaeology, and Environmental Studies, University of the Witwatersrand, Private Bag 3 WITS, 2050, South Africa

© 2020 Society for Ecological Restoration

doi: 10.1111/rec.13292

2 of 18 Restoration Ecology April 2021

 Research agenda for tropical grassland restoration

Figure 1. We identified emerging questions across broad themes in restoration (left of the figure) to structure questions aimed at optimizing practical restoration
outcomes for TGB (tropical grassy biomes, i.e. old-growth tropical and subtropical grasslands and savannas). Research should focus on: disentangling the reasons
why TGB are often undervalued and misunderstood; mapping TGB restoration opportunities; identifying and recognizing regions where alternative stable states
exist (a—Suding et al. 2004) and recognizing areas with natural regeneration potential to avoid unnecessary intervention; restoring abiotic conditions;
disentangling factors driving low seed quality; determining germination requirements and developing vegetative propagation techniques for TGB species;
disentangling the limiting factors and key ecological processes underlying seedling establishment and community assembly; improving and validating long-term
management to mimic natural disturbance regimes; and improving monitoring.

and importance of the ecosystem services provided by TGB types, information that is not readily available in many trop-
(Bengtsson et al. 2019), which can include, but should not be ical and subtropical regions. Ideally, mapping the global and
limited to, carbon sequestration, which is the standard approach regional distribution of grasslands should be aligned with
used in forested landscape restoration initiatives. mapping the distribution of forests, thus avoiding the kinds
of overlap pointed out by Veldman et al. (2015). Careful
planning for TGB restoration includes proper identification
Defining TGB Restoration Opportunities and Targets of: (1) pristine TGB as reference ecosystems; (2) degraded
Restoration Opportunities TGB which constitute restoration targets; and (3) degraded
forests converted to derived grasslands and savannas which
Identifying restoration opportunities for TGB requires: (1) infor-
constitute reforestation targets. Mapping these three ecosys-
mation on the distribution of TGB, as well as degraded and
tem types using remote sensing data derived from satellites
derived grasslands at different scales, (2) finer-scale information
should provide important information for regional and
on the distribution of TGB, and (3) potential socioeconomic
global analyses. However, because of their temporal and
constraints to feasibility, all of which can be mapped. Setting
spatial resolution limitations, remote sensing techniques
restoration priorities entails valuation of restoration benefits to
may not be precise enough and fail to distinguish TGB from
society and using scientific evidence to make appropriate,
degraded and derived grasslands (Veldman & Putz 2011).
informed, and transparent prioritization decisions.
The recent Global Ecosystem Dynamics Investigation
(GEDI) provides a powerful option to identify the complex-
(1) A critical step to adequately plan restoration is to have reli- ity of TGB. First, it uses a high-resolution laser that allows
able information on the distribution of different grassland quantification of the vertical distribution of vegetation and

April 2021 Restoration Ecology 3 of 18

Research agenda for tropical grassland restoration

Figure 2. Pristine TGB are sometimes confused with degraded grasslands: (A) open savannas in the Cerrado (TGB), Parque Estadual do Jalap~ao, Tocantins,
Brazil, composed of a continuous species-rich herbaceous layer and scattered shrubs and small trees; the photo also shows two rhea (Rhea americana), a bird
native to these ecosystems (photo credit AF); (B) a former Pinus spp. plantation in the Cerrado after removal of the trees. Melinis minutiflora (purple
inflorescence), an African grass, has invaded the system, outcompeting native grasses and forbs (photo credit AF); (C) native grassland (TGB) in Serra dos
Carajás, Pará, Brazil (photo credit AJA); (D) derived grassland replacing former forest in Carajás National Forest, Pará, Brazil (photo credit AJA); (E) open
Acacia savanna (TGB), with a continuous herbaceous layer and scattered trees and shrubs, during the wet season in Kruger National Park, South Africa; the photo
also shows ostriches (Struthio camelus), which are native to these systems (photo credit DT); (F) open Acacia savanna (TGB) during the dry season, Kruger
National Park, South Africa (photo credit DT).

estimates variables including surface topography, canopy global map of TGB restoration opportunities. Large-scale
height and cover metrics, and vertical structure metrics mapping should also incorporate spatial autocorrelation
(Dubayah et al. 2020). Second, GEDI-derived data can be and spatial structure data to improve the predictive ability
downloaded free. The feasibility of using both vertical and of mapping opportunities for restoration at the global scale
horizontal information to facilitate floristic differentiation (Ploton et al. 2020).
and allow the separation of TGB from degraded and derived (2) In many regions, TGB are composed of multiple grass-
grasslands needs to be empirically tested. Remote sensing dominated ecosystems found in mosaics including woody
assessments should be coupled with in situ biodiversity vegetation types (Sankaran et al. 2005): (1) forming a con-
assessments to calibrate the models and produce a reliable tinuum from grassland and savanna to closed-canopy

4 of 18 Restoration Ecology April 2021

 Research agenda for tropical grassland restoration

woodland and forests; (2) representing mosaics of natural Developing Techniques to Restore Soil Conditions
alternative stable states in the landscape (Bond 2019); When degradation entails total or partial loss of desirable soil
and (3) characterizing patterns due to edaphic factors or dis- characteristics, earthwork may constitute the first restoration
turbance (Buisson et al. 2019; Le Stradic et al. 2015a; step, as in cases of post-mining restoration. While reprofiling
Fig. 3). Each grassland type should be restored where it mine sites is often done (Fig. 4), restoring hydrological pro-
belongs, recreating mosaics to restore biodiversity and ser- cesses (including restoring soil infiltration capacity or reducing
vices at the landscape scale, and maintained using manage- surface runoff), soil structure (the degree of aggregate stability),
ment or natural disturbances. chemical properties, and microbial ecology (decomposition and
(3) Restoration opportunities are often constrained by environ- nutrient cycling) are poorly studied for TGB.
mental legislation, social will, and available resources to Soil aggregation is the result of soil biota and their organic
implement restoration. Therefore, socioecological benefits products, and a driver of soil water movement and retention, ero-
should also be mapped and integrated in the definition of sion, crusting, nutrient recycling, root penetration, and produc-
restoration opportunities (Brancalion et al. 2019). Funds tivity (Bronick & Lal 2005). Soil aggregate stability is often
for landscape-scale restoration have increased more rapidly degraded by management or restoration interventions, such as
for forest than for nonforest ecosystems (Overbeck soil tillage (Plante & McGill 2002), so understanding how to
et al. 2015; Veldman et al. 2015). Since fund availability restore soil aggregation in TGB, aiming at preserving both soil
changes quickly in space and time in unpredictable ways, biota and hydrological processes, is key to improve restoration
frequent updating of this information is required. success.
For less extreme degradations, soil preparation may be the
first step in restoration. Tilling may efficiently break sealed soil
surfaces (Kinyua et al. 2010; Fig. 5) and topsoil removal may
Setting Restoration Targets reduce the seed bank of undesirable species (Pilon et al. 2018;
Identifying pristine ecosystems as references is challenging for Sampaio et al. 2019). Creating microtopography can contribute
TGB restoration because (1) they are old-growth ecosystems to reduce soil erosion and increase the capture of soil and seeds
and their restoration is often unachievable on human time- from the surrounding landscape (Kimiti et al. 2017). Overall,
scales (Veldman et al. 2015); (2) pristine ecosystems of a par- however, benefits and potential negative consequences of soil
ticular region may no longer exist; and (3) the highly dynamic preparation have seldom been studied in TGB. Little is known
nature of ecosystems (Higgs et al. 2018), particularly in about the effect of these practices on biological soil crusts. Tech-
response to periodic disturbance, hinder simple restoration tar- niques should be tested more widely, diversified (e.g. soil inver-
gets. Nonetheless, identifying a reference ecosystem sion), and compared. Many TGB species have a bud bank or
(or preferably a regional range) is critical in restoration plan- underground storage organs that confer resilience to above-
ning. In many situations, finding a reference ecosystem is chal- ground endogenous disturbances such as fire and herbivory
lenging where alternative stable states naturally exist, creating (Fig. 6). Therefore, disrupting the soil is advised only if the deg-
the possibility of restoring any of a range of grassland- radation has already largely eliminated the underground compo-
savanna-forest transition states in the landscape (Bond 2019). nent of the vegetation (Buisson et al. 2019). If the soil has not
Setting restoration goals for a particular project also requires been disturbed or compacted, soil preparation is likely to be inap-
discernment between what is desirable and what is possible propriate, as it will hamper recovery from the seed and bud bank.
under the local constraints and the needs of local communities Fertilization degrades many TGB on nutrient-poor soils (Harpole
(Ehrenfeld 2000). Ideally, restored ecosystems should be et al. 2016) and techniques to reduce soil nutrients, such as mow-
structurally similar to the reference ecosystem, be composed ing and exporting biomass, carbon amendment, and grazing
by a subset of species from the regional pool, and have suffi- (Sitters et al. 2020) warrant further exploration. In sum, restoring
cient functional diversity to sustain ecosystem functioning. It the physical and chemical properties of soils may be the key first
is crucial for these restored TGB ecosystems to be resilient to step to restore an important ecological filter that impacts commu-
endogenous disturbances (i.e. disturbances that are part of the nity assembly and desired restoration goals such as resilience and
evolutionary history and internal dynamics of a system). Opti- ecosystem functioning (Fig. 5).
mizing goal-setting for TGB restoration relies on multidisci-
plinary research aimed at: (1) identifying the preexisting
vegetation type under the same climatic conditions; (2) recog- Reintroducing Native Species From Seed
nizing potential alternative stable states and disrupting inter- Extensive knowledge gaps in basic seed biology of TGB species
nal feedbacks to allow the transitions from degraded to the severely limit our capacity to use seed-based restoration strate-
desirable states (Suding et al. 2004); (3) diagnosing the poten- gies. Landscape-scale restoration requires large quantities of
tial for natural regeneration before deciding to intervene seeds, but many TGB herbaceous species have traits that limit
(Zahawi et al. 2015; Meli et al. 2017) given the unclear rela- seed sourcing, including: (1) small populations, individual rar-
tive (cost-)effectiveness of passive versus active restoration ity, and irregular fruiting phenology (Dayrell et al. 2016);
in TGB; and (4) defining fundamental attributes for the TGB (2) low resource allocation to reproduction (Bond &
to be restored. Midgley 2001); (3) recalcitrant seeds in forbs that impede seed

April 2021 Restoration Ecology 5 of 18

Research agenda for tropical grassland restoration

Figure 3. In many regions, grass-dominated ecosystems have existed historically as alternative states in mosaics alongside woody vegetation types, often
forming a continuum from grassland and savanna to closed-canopy woodland and forest in relation to different disturbance regimes and/or environmental
~ a woodland, all from the Itirapina Ecological Station, SP, Brazil (photo credit GD); (D) grassland-Protea
gradients; (A) grassland, (B) savanna, (C) Cerradao,
savanna continuum in Marakele National Park, Waterberg, South Africa (photo credit DT). Grassland or savanna and forest can also represent alternative stable
states in the landscape: (E) grassland and dense savanna in Mole National Park, Ghana (photo credit GD); (F) mosaic of grassland and savanna in Magaliesberg,
South Africa (photo credit SJS). Mosaics of grassy and woody vegetation can also represent patterns due to edaphic factors: (G) grasslands and rocky outcrops in
Swaziland (photo credit SJS); (H) grassland-forest alternative stable states, here maintained by edaphic factors in Soutpansberg, South Africa (photo credit DT);
(I) mosaic of grassland, savanna, and forest in the Katanga Copperbelt, Democratic Republic of Congo (photo credit copperflora.org); (J) stony and sandy
grasslands, as well as rocky outcrops, in Serra do Cipó, Minas Gerais, Brazil (photo credit EB).
6 of 18 Restoration Ecology April 2021
 Research agenda for tropical grassland restoration

Figure 4. Large-scale earthwork often constitutes the first step in post-mining restoration. In the Katanga copperbelt, Democratic Republic of Congo, various
grassland communities occur on outcrops depending on a gradient of copper (Cu) and cobalt (Co) concentrations. The image shows a post-mining restoration site
where a hill was reshaped to create the template for restored grassland communities. The hill comprises relatively high Cu–Co concentration subsoil. A small area
of high Cu–Co concentration topsoil was spread on which vegetation has already established (photo credit S Le Stradic, Technische Universität München,
Germany).

storage; (4) high seed dormancy that inhibits germination create opportunities to improve sourcing seeds from native TGB
(Rusdy 2017); (5) low seed quality (low viability, high percent- species where fire-stimulated flowering typically results in higher
age of empty seeds, Dairel & Fidelis 2020; Kolb et al. 2016; and more synchronous seed production and higher seed quality
Dayrell et al. 2017); and (6) fire-dependent flowering, seed (Young 2004; Le Stradic et al. 2015b; Wagenius et al. 2020).
release, or germination (Bond 2019). Seed germination of her- Direct seeding for large-scale ecosystem restoration can be a
baceous species is often low, slow, and asynchronous. Therefore cost-effective technique to reintroduce functionally diverse
determining germination requirements (including fire-related plant communities (Fig. 7). However, it has rarely been imple-
cues), dormancy classes for TGB species, and methods to over- mented in TGB because of the overall low seed quality men-
come dormancy, will be useful to optimize seedling establish- tioned above and because native (local) seeds are not available
ment and decrease costs, increase the number of herbaceous in sufficient quantities. Two pioneering studies tested direct
species included in restoration programs (Ladouceur et al. 2018), seeding for TGB restoration, but used relatively few grass and
and avoid scenarios where shorter-lived species are preferred forb species compared to local diversity (13 grass species in
over long-lived perennials because of a shortage of viable seeds Waters et al. 2001; 11 grass and 3 forb species in Sampaio
with known germination techniques. et al. 2019). While these studies are necessary steps toward effi-
Putative causes driving low seed quality in TGB species include cient seeding, more research on seed ecology is urgently needed.
nutrient-limitation (Fujita et al. 2014), pollen quality (Rosbakh Conversely, the high demand for seeds may have positive
et al. 2018), higher genetic load due to accumulation of deleterious impacts, including (1) involvement of local communities, gener-
somatic mutations over many disturbance cycles (Lamont & ating income and the support of livelihoods through the sustain-
Wiens 2003), harvesting constraints when faced with irregular able use of native biodiversity; (2) protection of native
fruiting phenology, and a combination of these factors. It also ecosystems as seed sources; (3) decreasing conflict between
remains to be tested to what extent managing fire regimes would environmental agencies, local communities, and companies that

April 2021 Restoration Ecology 7 of 18

Research agenda for tropical grassland restoration

Figure 5. Soil preparation may be the first step in restoration: (A) ripping allows the breaking up of sealed soil surfaces, leading to natural recruitment by Cynodon
sp., Kenya (photo credit TPY); (B) soil plowing before direct seeding in Chapada dos Veadeiros National Park, Goiás, Brazil; many TGB species have
underground storage organs that confer resilience to endogenous disturbance (Fig. 6), so disrupting the soil is advised only if the degradation has already largely
destroyed the underground component of the vegetation; benefits and potential negative consequences of soil preparation have seldom been studied (photo credit
IS); (C) and (D) soil preparation for seeding with Themeda triandra in gel, which provides sustained moisture for germination, South Africa (photo credits KK).

Figure 6. Non-graminoid TGB species often have underground storage organs or a bud bank that confers resilience: (A) Commicarpus pentandrus, South Africa
(photo credit SJS); (B) Gomphrena macrocephala, Itirapina Ecological Station, SP, Brazil (photo credit EB); (C) Crotalaria cornetii, Katanga, Democratic
Republic of Congo (photo credit copperflora.org).

8 of 18 Restoration Ecology April 2021

 Research agenda for tropical grassland restoration

Figure 7. (A) Direct sowing of native seeds on an opencast coal mine, South Africa. Seeds were harvested from undisturbed grasslands within 20 km of the
rehabilitation site, using hand-held flail harvesters, which collected seeds from all grass and forb species present. Seeds were not separated or sorted, and were air-
dried and stored for 1 year, thereafter they were mixed into a gel and placed in the soil, in the same species ratios as harvested and (B) the same site in South Africa
4 years later; the dominant grass species established were Eragrostis curvula, Eragrostis racemosa, Melinis repens, Paspalum scrobiculatum, Perotis patens,
Pogonarthria squarrosa, and Themeda triandra (photo credit KK); (C) mechanized direct seeding of grasses, forbs, shrubs, and trees in degraded Neotropical
savanna in Central Brazil; and (D) broadcast seeding of these same life-forms performed by volunteers in Chapada dos Veadeiros National Park, Goiás, Brazil
(photo credits F Tatagiba, ICMBio).

need to perform restoration projects (Schmidt et al. 2019b); and Reintroducing Native Species by Vegetative
(4) greater cost-effectiveness associated with greater market Propagation and Container Stock
demand. Because seeds are currently sourced from wild popula- A large proportion of plant species from TGB have at least one of
tions, thresholds of seed collection need to also be determined. the abovementioned traits that limit propagation from seeds.
Seed provision/supply chains must be supported and regulated, Therefore, the opportunities for their reintroduction and propaga-
for which preliminary law and market analysis research are nec- tion via seed sowing, soil transfer, and hay transfer are currently
essary (de Urzedo et al. 2019). Although harvesting seeds from often limited for restoring TGB biodiversity (Le Stradic
perennial species tends to have low impacts on adult plants and et al. 2015b; Dayrell et al. 2016; Kolb et al. 2016; Pilon
populations (Franco & Silvertown 2004), clear harvesting et al. 2018). To be reintroduced, such species may depend mainly
guidelines are needed to avoid or minimize negative impacts on vegetative propagation. Cerrado grassland restoration has
to plant populations, especially when these are conservation tar- proven to be successful at a small scale by transplanting grass
gets. Economic stimulus for the development of seed farming tussocks when there are sources available (Pilon et al. 2019).
production can alleviate shortfalls in seed supply, improving However, for many TGB the transplantation of forbs, subshrubs
the reliability and cost of seeds available for restoration and and shrubs, sometimes even grasses (Le Stradic 2012), is not a
attenuating the potential for overharvesting of wild plant popu- promising technique. Alternative techniques must then be devel-
lations. Seed farming should address evolutionary selection oped or applied, such as: transplantation of individual tillers
and loss of genetic diversity (Broadhurst et al. 2008, 2015; grown in pots from fragmented tufts (du Toit 2009), transplanta-
Nevill et al. 2016). However, this also implies the development tion of seed-germinated pot-grown grasses (de Oliveira
of Seed Transfer Zones, as already developed for some temper- et al. 2020), cuttings (du Toit 2009; Martins et al. 2015), propa-
ate regions, e.g. in Germany (Durka et al. 2017). Given the large gation from bulbs (S. Le Stradic, Technische Universität
knowledge gaps on TGB composition and regional differentia- München, Germany) or roots (Silva et al. 2011), and micropropa-
tion, the development of ecologically meaningful transfer zones gation (Biondo et al. 2007; de Oliveira et al. 2011; de Souza
will still require considerable research efforts for most regions. et al. 2015), among others (Fig. 8).

April 2021 Restoration Ecology 9 of 18

Research agenda for tropical grassland restoration

Figure 8. Reintroduction of recalcitrant species may depend mainly on vegetative propagation. Research should focus on propagating TGB plants from stem
cuttings, bulbs, or roots via micropropagation: (A) identification of the niche of four populations of Aeollanthus saxatilis and propagation trials from tubers, in
Katanga, Democratic Republic of Congo (photo credit copperflora.org); (B) in vitro cultivation trial of Rhynchospora consanguinea from campo rupestre
grasslands, Cerrado, Brazil (photo credit EB); (C) large-scale restoration of Cerrado grassland by transplanting, in Santa Bárbara State Forest, SP, Brazil (photo
credit GD); and (D) the same site 3 years after transplanting (photo credit GD).

While vegetative propagation has been widely used for for- vegetative propagation, direct seeding, and container stock.
estry and orcharding worldwide, it has rarely been mentioned Examining the costs and benefits of each approach will depend
in the context of ecological restoration, and its application at on restoration goals and the socioecological context of each
large scale relies on research and technological development, grassland.
while assuring high genetic variability. Protocols for the rein-
troduction of TGB species from direct rescue transplantation,
or plug production from cuttings, division, scarce seed quanti- Maximizing Plant Establishment and Ecosystem
ties, or by micropropagation are needed (Koch 2007). Develop- Sustainability
ment of propagation protocols should take advantage of the Poor seedling establishment in TGB is often related to seed qual-
underground storage organs and bud banks which are reliable ity (see above), competition from undesirable species, severe
sources of regeneration in TGB species (Fidelis et al. 2014), environmental filters acting upon seedling survival, or a combi-
and which hold the promise of simultaneously increasing vege- nation of these. Research needs to focus on improving (1) resto-
tation resilience and providing ecosystem services (Pausas ration treatments to maximize species establishment; and
et al. 2018). Most likely, solutions for successful TGB restora- (2) subsequent long-term management promoting ecosystem
tion will be based on a mix of different approaches including functioning and sustainability.

10 of 18 Restoration Ecology April 2021

 Research agenda for tropical grassland restoration

Maximizing Target Species Establishment disturbance regime and optimum treatments to keep a balance
Controlling invasive herbaceous species (including intention- between the establishment of native and invasive species is yet
ally introduced forage species), and invasive or native tree and to be determined through long-term field experiments in both
shrub encroachment is a major restoration step to reduce compe- disturbance-dependent and edaphic TGB (sensu Buisson
tition from undesirable species. Numerous techniques are avail- et al. 2019).
able including cutting, removal, grazing, prescribed fires Community assembly in old-growth grasslands occurs at
(Fig. 9), herbicide application, or, as mentioned in the section large temporal scales (Nerlekar & Veldman 2020), so pioneer
“Developing Techniques to Restore Soil conditions,” deep or faster-growing species which are typically used in the early
plowing or topsoil removal if the site is already highly degraded stages of forest or temperate grassland restoration (Silveira
(Buisson et al. 2019). The most suitable combination of treat- et al. 2020) are either unknown, or are unsuitable for TGB
ments is, however, yet to be determined for each TGB type restoration. Some studies identify and employ faster-growing,
and for each invasive species. short-lived species to cover the ground and later allow the
Many TGB are maintained by specific fire regimes and graz- establishment of long-lived perennial species (e.g. Coutinho
ing (Veldman et al. 2015). While reintroduction of fire or graz- et al. 2019). However, these species might not guarantee the
ing regimes too early may hamper the establishment of resilience of the system because they may actually inhibit the
reintroduced species, excluding them for excessive periods subsequent establishment of perennial species (Nerlekar &
decreases diversity, promotes woody encroachment and biolog- Veldman 2020). While introducing pioneer species could be
ical invasion, and increases the danger of wildfire occurrence, all useful in areas where invasive species and soil erosion are
of which damage the system (Stevens et al. 2017; Alvarado major issues, the ability of this practice to promote high biodi-
et al. 2018; Fidelis et al. 2018). Therefore, the appropriate versity in the long-term remains to be demonstrated.

Figure 9. Prescribed fires can be used to positive effect in the restoration process: (A) study of the effect of fire to control the invasive African grass Urochloa
brizantha (syn. Brachiaria brizantha) invading the Cerrado at the Itirapina Ecological Station, SP, Brazil (photo credit AF); (B) study of the effect of fire to
control the grass Melinis minutiflora invading the Cerrado; the image shows the plant community after a prescribed dry season burn conducted in July 2014. The
photo was taken in October 2015, at the beginning of the rainy season; M. minutiflora has germinated or resprouted, along with some native species (e.g. Eugenia
punicifolia at the top of the plot, with some Cyperaceae below, Qualea grandiflora and Pradosia brevipes to the left and right of the plot, respectively), Itirapina
Ecological Station, SP, Brazil (photo credit G Damasceno, UNESP Rio Claro); (C) study of the role of fire in removing pine litter and stimulating native species
recovery after degradation of coastal grasslands by tree plantations, Lagoa do Peixe National Park, Rio Grande do Sul, Brazil; and (D) study of the role of fire in
removing pine litter and stimulating native species recovery after degradation by tree plantations in highland grasslands in southern Brazil. One year after the
experimental burn vegetation is dominated by native ruderal sedges from the soil seed bank, while grasses typical of the reference systems did not recruit,
indicating the need for active seed introduction (photo credits GEO).

April 2021 Restoration Ecology 11 of 18

Research agenda for tropical grassland restoration

Alternatively, restoring small, slow-growing species to facili- mycorrhizal fungi from the native plant microbiome in temper-
tate colonization by other native species, and to thus increase ate grassy ecosystems (Koziol et al. 2018).
native species richness can be a feasible strategy (Fynn
et al. 2009). TGB restoration may benefit from trait-based
approaches for grasses and forbs to address the functional role Maximizing Ecosystem Sustainability (Long-Term Management)
of herbaceous species in promoting community reassembly When restoration goals include achieving a sustainable ecosys-
(Silveira et al. 2020). tem, this means ensuring that (1) degradation forces remain
We also need a better understanding of feedbacks between excluded; (2) introduced plants persist in the longterm; (3) addi-
vegetation, fauna, soil microbiota, and soil properties (Suding tional native species can establish and successfully reproduce;
et al. 2004). Mycorrhiza, soil microbes (Lugo & Pagano 2019), and (4) restored ecological processes maintain ecosystem func-
ants (Parr et al. 2016), large herbivores, including livestock tioning (Miller et al. 2017). In disturbance-dependent TGB,
(Veblen 2012; Hempson et al. 2015; Fritz 2017), and nurse wildfires and large herbivores play underappreciated roles not
plants (Gómez-Aparicio 2009) have been shown to play impor- only in restoration (see section “Maximizing Target Species
tant roles in enhancing target species establishment and persis- Establishment”), but also in maintaining long-term sustainabil-
tence, but we still lack proper identification of key ecological ity and resilience (Asner et al. 2009; Burkepile et al. 2013; Rigi-
engineers with the potential to maximize plant establishment nos et al. 2018).
in most TGB. Examining the role of microorganisms in the Maintaining an appropriate level of herbivory may entail
establishment of TGB species is worthy given recent evidence either of two very different groups of large mammalian herbi-
that restoration outcomes can be improved by reintroducing vores: native herbivores and domestic livestock (Fig. 10).

Figure 10. Maintaining an appropriate level of herbivory may entail either of two different groups of large mammalian herbivores, being native herbivores and
domestic livestock: (A) native herbivores and domestic livestock coexisting, Kenya (photo credit RM Pringle, Princeton, NJ); (B) foraging behavior and carrying
capacities, e.g. for the African elephant Loxodonta africana, must be studied to carefully plan the restoration of diverse herbivore assemblages in savannas such as
the Kruger National Park, South Africa (photo credit DT); (C) understanding the synergistic effects of fire and herbivory on structuring the herbaceous vegetation
component is critical in savanna restoration and management. Here, mixed herbivores (giraffe Giraffa camelopardalis and Burchell’s zebra Equus quagga
burchellii) are excluded from areas of fire-manipulated savanna, Kruger National Park, South Africa (photo credit DT); and (D) blue wildebeest Connochaetes
taurinus grazing in and around an exclosure designed to preclude only the largest of the African herbivores (photo credit D Burkepile, University of California
Santa Barbara).

12 of 18 Restoration Ecology April 2021

 Research agenda for tropical grassland restoration

Native herbivores (and perhaps livestock, see below) not only Thus, addressing restoration of diverse native herbivore assem-
have an important role in shaping TGB vegetation species com- blages (refaunation; Seddon et al. 2014) to target sites may be
position and structural diversity via their foraging, but also for critical to achieving sustainable TGB goals.
distributing and maintaining nutrient inputs via dung deposits Domestic livestock in TGB are represented by a diverse set of
(Fornara & Du Toit 2008; Asner et al. 2009; Sitters et al. 2020). species (cattle, sheep, goats, camels, donkeys) that, like wildlife,

Figure 11. Maintaining fire in TGB is important, but the frequency, timing, and pattern of the fire regime must be studied: (A) fire experiments in wet grasslands
in Jalap~ao, Tocantins, Brazil (photo credit AF); (B) fire team measuring environmental conditions during fire experiments in Reserva Natural Serra do Tombador,
Goiás, Brazil (photo credit AF); (C) early dry season (July) prescribed fire and (D) late dry season (October) prescribed fire, both in Reserva natural Serra do
Tombador, Goiás, Brazil (photo credit AF); (E) endangered woodbush granite grassland degraded through encroachment by woody species, Haenertsburg Nature
Reserve, South Africa and (F) prescribed fire being used to remove encroaching woody species and restore biodiversity in the same grassland (photo credit DT);
(G) fire frequency experiments have been carried out for over 60 years on experimental plots in the Kruger National Park, South Africa (photo credit D Burkepile,
University of California Santa Barbara).

April 2021 Restoration Ecology 13 of 18

Research agenda for tropical grassland restoration

differ in dietary preferences and ecosystem impacts. Livestock is Fires have declined worldwide in recent decades (Andela
often the underlying cause of ecosystem degradation, through et al. 2017), but intentional fires for grassland management are
inappropriately high stocking densities. However, at moderate still common in parts of South America (Eloy et al. 2019;
densities, it may effectively replace native diverse herbivore Schmidt & Eloy 2020) and Madagascar (Alvarado et al. 2018)
communities (Veblen et al. 2016). This may be especially and eastern and southern Africa (Archibald 2016). Even there,
important in TGB where the native megafauna has been lost increased grazing and bush encroachment (both of which reduce
and cannot be reintroduced or where prescribed fires are not per- grass fuels) have locally decreased burning (Archibald 2016). In
mitted. Yet, the social and economic forces that led to overgraz- Kenya, grassland fires have become rare (Archibald 2016) both
ing are often still in place, and not easy to shift (Kimiti because of heavy grazing in communal rangelands and because
et al. 2017). Specific studies thus need to determine appropriate private range managers are increasingly disinclined to burn live-
overall grazing rates, and spatial and temporal patterns that allow stock forage (R. Sensenig, Goshen College, IN, personal com-
the maintenance of habitat heterogeneity and biodiversity. Study- munication). In the latter case, fire reintroduction may require
ing which levels or combinations of domestic and wild native stakeholder outreach or environmental sociology studies. The
herbivore grazing are appropriate is an essential issue to allow strong synergistic effects of fire and herbivory, which can make
restoration and sustainable use of TGB (Riginos et al. 2012). the combination a far more powerful force than either alone
Many TGB are maintained by fire (Bond 2019), suggesting (Midgley et al. 2010; Werner et al. in press), have been investi-
that fire management may be needed for restoration (but see gated for woody plants (Staver et al. 2009; Pringle et al. 2015),
edaphic grasslands; Buisson et al. 2019). However, understand- but remain understudied for the herbaceous component of
ing natural fire regimes is challenging because of entrenched TGB (but see Burkepile et al. 2016; Smith et al. 2016).
anthropogenic fire regimes often preventing an accurate assess-
ment of natural fire frequency, timing, and patterns (Archibald
et al. 2013; Johansson et al. 2020; Fig. 11). Maintaining a mosaic Monitoring
of diverse fire regimes may be a way to maintain landscape het- Monitoring is essential to evaluate restoration success and is par-
erogeneity and diversity, but application of specific fire regimes ticularly important for restored disturbance-dependent ecosys-
in support of restoration should be further studied (Bond 2019). tems that require long-term adaptive management (Holl &

Figure 12. Sampling species diversity and functional traits in pristine TGB reference ecosystems is critical in establishing the benchmark for local restoration
goals, Haenertsburg Nature Reserve, South Africa (photo credit S Kremer-Kӧhne, Friends of the Haenertsburg Grasslands).

14 of 18 Restoration Ecology April 2021

 Research agenda for tropical grassland restoration

Cairns Jr. 2002). While monitoring indicators such as tree cover, see Dey et al. 2020), ultimately resulting in long-term benefits
aboveground biomass, and species richness may be appropriate to people and nature in some of the most neglected ecosystems
for forest restoration, this alone is not sufficient to determine of our planet.
TGB health (Silveira et al. 2020). The suite of indicators to be
monitored depends on the restoration goals (Holl & Cairns
Jr. 2002) and on clear definitions of the expected attributes of Acknowledgments
the desirable ecosystem. This should be better explored for
TGB. Candidate indicators for TGB monitoring can be general, E.B., A.F., A.J.A., F.A.O.S. are supported by CNRS PICS
such as ground cover or species richness, above and below- 2018–2020 [RESIGRASS], E.B. and T.P.Y. by France-
ground biomass, proportion between different life-forms, fidel- Berkeley Fund [https://fbf.berkeley.edu/], F.A.O.S., G.D.,
ity to reference community, abundance of indicator species, R.S.O., G.E.O., and A.F. are supported by grants from CNPq,
F.A.O.S. by FAPEMIG, T.P.Y. by NSF DEB 10-50543, and
among others, but should also include indicators specific to deg-
LTREB 19-31224, G.E.O. by Fundaç~ao O Boticário, A.F. by
radation and resilience to disturbance, such as recovery after fire,
drought, or herbivory, presence and cover by ruderals and exotic FAPESP (2015/06743-0; 2019/09903-0), R.S.O. by FAPESP-
species, soil erosion rates, soil compaction, or woody NERC (2019/07773-1), M.H.S. by the Claude Leon Founda-
encroachment. tion. G.E.O., A.F., G.D. participate in GrassSyn within the SIN-
Ecosystem functioning is easily affected by changes in the BIOSE/CNPq framework (grant 42348/2019-3 to G.E.O.). The
authors thank two anonymous reviewers and Lars Brudvig for
suite of functional traits present. Monitoring native TGB species
their comments, which significantly improved the manuscript.
richness alone may infer that a TGB has been restored ade-
quately, although functional richness, and hence redundancy,
may have collapsed, which will negatively affect the resilience
of the restored ecosystem over time. Monitoring of functional LITERATURE CITED
diversity in TGB is therefore imperative to properly evaluate Alvarado ST, Silva TSF, Archibald S (2018) Management impacts on fire occur-
restoration success, providing that the traits to be assessed have rence: a comparison of fire regimes of African and South American tropical
been shown to be relevant to TGB functioning and their ecosys- savannas in different protected areas. Journal of Environmental Manage-
tem services (Fig. 12). Research aimed at developing indicators ment 218:79–87
Andela N, Morton DC, Giglio L, Chen Y, van der Werf GR, Kasibhatla PS, et al.
of TGB functioning should identify easy-to-measure surrogates
(2017) A human-driven decline in global burned area. Science 356:
for ecosystem attributes and processes that are difficult to assess 1356–1362
(Dale & Beyeler 2001; Prach et al. 2019). Because a number of Anderegg WRL, Trugman AT, Badgley G, Anderson CM, Bartuska A, Ciais P,
ecological processes (e.g. regeneration, succession and et al. (2020) Climate-driven risks to the climate mitigation potential of for-
landscape-level connectivity) operate differently in open eco- ests. Science 368:eaaz7005
systems than they do in forests (Bond 2019; Silveira et al. 2020), Archibald S (2016) Managing the human component of fire regimes: lessons from
Africa. Philosophical Transactions of the Royal Society, B: Biological Sci-
monitoring TGB restoration is likely to span comparatively
ences 371:20150346
larger time frames than in forest restoration. This assumption Archibald S, Lehmann CER, Gómez-Dans JL, Bradstock RA (2013) Defining
remains to be tested across a range of TGB. pyromes and global syndromes of fire regimes. Proceedings of the National
Academy of Sciences 110:6442–6447
Asner GP, Levick SR, Kennedy-Bowdoin T, Knapp DE, Emerson R, Jacobson J,
Conclusion Colgan MS, Martin RE (2009) Large-scale impacts of herbivores on the
The rate of land use conversion outpaces the rate of scientific structural diversity of African savannas. Proceedings of the National Acad-
emy of Sciences 106:4947–4952
knowledge production, posing a great threat to the long-term
Bengtsson J, Bullock JM, Egoh B, Everson C, Everson T, O’Connor T,
persistence of TGB and thus calling for collaborative research O’Farrell PJ, Smith HG, Lindborg R (2019) Grasslands—more important
prioritization. Further, this escalating transformation makes the for ecosystem services than you might think. Ecosphere 10:e02582
need for effective restoration initiatives all the more pressing. Biondo R, Souza AV, Bertoni BW, Soares AM, França SC, Pereira AMS (2007)
Here, we have identified emerging questions on ecological Micropropagation, seed propagation and germplasm bank of Mandevilla
dimensions of restoration ecology (Perring et al. 2015) that need velutina (Mart.) Woodson. Scientia Agricola 64:263–268
to be answered in order to scale up restoration of TGB to meet Bond WJ (2019) Open ecosystems: ecology and evolution beyond the forest
edge. Oxford University Press, Oxford, New York, https://doi.org/10.
the challenges of the UN Restoration Decade. Achieving resto-
2989/10220119.2019.1695663
ration targets for TGB by 2030 seems unlikely because these Bond WJ, Midgley JJ (2001) Ecology of sprouting in woody plants: the persis-
megadiverse ecosystems do not even have a collective name to tence niche. Trends in Ecology & Evolution 16:45–51
describe their ecology (Bond 2019), let alone sufficient knowl- Brancalion PHS, Niamir A, Broadbent E, Crouzeilles R, Barros FSM,
edge to restore them. The set of priorities discussed here are Zambrano AMA, et al. (2019) Global restoration opportunities in tropical
not without bias, but has been assembled by a team of experts rainforest landscapes. Science Advances 5:eaav3223
Broadhurst LM, Lowe A, Coates DJ, Cunningham SA, McDonald M, Vesk PA,
facing challenges during TGB restoration for more than two
Yates C (2008) Seed supply for broadscale restoration: maximizing evolu-
decades. They constitute a first step toward a comprehensive tionary potential. Evolutionary Applications 1:587–597
effort to improve the success of TGB restoration worldwide Broadhurst L, Hopley T, Li L, Begley J (2015) Using seed production areas to
(Fig. 1). Providing answers to these questions has the potential meet restoration targets and secure genetic diversity. Journal of the
to advance theory, practice, and policy in TGB restoration (but Australian Network for Plant Conservation 23:7

April 2021 Restoration Ecology 15 of 18

Research agenda for tropical grassland restoration

Bronick CJ, Lal R (2005) Soil structure and management: a review. Geoderma Fritz H (2017) Long-term field studies of elephants: understanding the ecology
124:3–22 and conservation of a long-lived ecosystem engineer. Journal of Mammal-
Buisson E, Stradic SL, Silveira FAO, Durigan G, Overbeck GE, Fidelis A, et al. ogy 98:603–611
(2019) Resilience and restoration of tropical and subtropical grasslands, Fujita Y, Venterink HO, van Bodegom PM, Douma JC, Heil GW, Hölzel N, et al.
savannas, and grassy woodlands. Biological Reviews 94:590–609 (2014) Low investment in sexual reproduction threatens plants adapted to
Burkepile DE, Burns CE, Tambling CJ, Amendola E, Buis GM, Govender N, phosphorus limitation. Nature 505:82–86
Nelson V, Thompson DI, Zinn AD, Smith MD (2013) Habitat selection Fynn RW, Wragg PD, Morris CD, Kirkman KP, Naiken J (2009) Vegetative traits
by large herbivores in a southern African savanna: the relative roles of predict grass species’ invasiveness and the invasibility of restored grass-
bottom-up and top-down forces. Ecosphere 4:art139 land. African Journal of Range and Forage Science 26:59–68
Burkepile DE, Thompson DI, Fynn RWS, Koerner SE, Eby S, Govender N, et al. Gann GD, McDonald T, Walder B, Aronson J, Nelson CR, Jonson J, et al. (2019)
(2016) Fire frequency drives habitat selection by a diverse herbivore guild International principles and standards for the practice of ecological restora-
impacting top-down control of plant communities in an African savanna. tion. Second edition. Restoration Ecology 27:S1–S46
Oikos 125:1636–1646 Gómez-Aparicio L (2009) The role of plant interactions in the restoration of
CBD - Convention on Biological Diversity (2011) Aichi Biodiversity Targets. degraded ecosystems: a meta-analysis across life-forms and ecosystems.
https://www.cbd.int/sp/targets/ (accessed 14 Oct 2019) Journal of Ecology 97:1202–1214
Climate Summit (2014) Forests - action statements and action plans. United Harpole WS, Sullivan LL, Lind EM, Firn J, Adler PB, Borer ET, et al. (2016)
Nations - Climate Change, New York, https://unfccc.int/sites/default/files/ Addition of multiple limiting resources reduces grassland diversity. Nature
new-york-declaration-on-forests_26-nov-2015.pdf (accessed 29 Sep 2020) 537:93–96
Cooke SJ, Bennett JR, Jones HP (2019) We have a long way to go if we want to Hempson GP, Archibald S, Bond WJ, Ellis RP, Grant CC, Kruger FJ, et al.
realize the promise of the “Decade on Ecosystem Restoration.” Conserva- (2015) Ecology of grazing lawns in Africa. Biological Reviews 90:
tion Science and Practice 1:e129 979–994
Coutinho AG, Alves M, Sampaio AB, Schmidt IB, Vieira DLM (2019) Effects of Higgs ES, Harris JA, Heger T, Hobbs RJ, Murphy SD, Suding KN (2018) Keep
initial functional-group composition on assembly trajectory in savanna res- ecological restoration open and flexible. Nature Ecology & Evolution 2:
toration. Applied Vegetation Science 22:61–70 580–580
Dairel M, Fidelis A (2020) How does fire affect germination of grasses in the Cer- Holl KD, Cairns J Jr (2002) Monitoring and appraisal. Pages 411–425.
rado? Seed Science Research, https://doi.org/10.1017/S0960258520000094 In: Perrow MR, Davy AJ (eds) Handbook of restoration ecology - 2. Resto-
Dale VH, Beyeler SC (2001) Challenges in the development and use of ecological ration in practice. Cambridge University Press, Cambridge, UK; New York
indicators. Ecological Indicators 1:3–10 IRP - International Resource Panel (2019) Global resources outlook 2019: natural
Dayrell RLC, Arruda AJ, Buisson E, Silveira FAO (2016) Overcoming chal- resources for the future we want. United Nations - Environment Pro-
lenges on using native seeds for restoration of megadiverse resource-poor gramme, Nairobi, Kenya, https://www.resourcepanel.org/reports/global-
environments: a reply to Madsen et al. Restoration Ecology 24:710–713 resources-outlook (accessed 29 Sep 2020)
Dayrell RLC, Garcia QS, Negreiros D, Baskin CC, Baskin JM, Silveira FAO IUCN-DC (2011) The Bonn challenge, http://www.bonnchallenge.org/content/
(2017) Phylogeny strongly drives seed dormancy and quality in a climati- challenge (accessed 14 Oct 2019)
cally buffered hotspot for plant endemism. Annals of Botany 119:267–277 Johansson MU, Abebe FB, Nemomissa S, Bekele T, Hylander K (2020) Ecosys-
De Groot RS, Blignaut J, Ploeg SVD, Aronson J, Elmqvist T, Farley J (2013) tem restoration in fire-managed savanna woodlands: effects on biodiver-
Benefits of investing in ecosystem restoration. Conservation Biology 27: sity, local livelihoods and fire intensity. Ambio. https://doi.org/10.1007/
1286–1293 s13280-020-01343-7
Dey CJ, Rego AI, Midwood JD, Koops MA (2020) A review and meta-analysis Kimiti DW, Riginos C, Belnap J (2017) Low-cost grass restoration using erosion
of collaborative research prioritization studies in ecology, biodiversity con- barriers in a degraded African rangeland. Restoration Ecology 25:376–384
servation and environmental science. Proceedings of the Royal Society B: Kinyua D, McGeoch LE, Georgiadis N, Young TP (2010) Short-term and long-
Biological Sciences 287:20200012 term effects of soil ripping, seeding, and fertilization on the restoration of
Dubayah R, Blair JB, Goetz S, Fatoyinbo L, Hansen M, Healey S, et al. (2020) The a tropical rangeland. Restoration Ecology 18:226–233
global ecosystem dynamics investigation: high-resolution laser ranging of Koch JM (2007) Alcoa’s mining and restoration process in South Western
the Earth’s forests and topography. Science of Remote Sensing 1:100002 Australia. Restoration Ecology 15:S11–S16
Durka W, Michalski SG, Berendzen KW, Bossdorf O, Bucharova A, Hermann J- Kolb RM, Pilon NAL, Durigan G, Kolb RM, Pilon NAL, Durigan G (2016) Fac-
M, Hölzel N, Kollmann J (2017) Genetic differentiation within multiple tors influencing seed germination in Cerrado grasses. Acta Botânica Brasí-
common grassland plants supports seed transfer zones for ecological resto- lica 30:87–92
ration. Journal of Applied Ecology 54:116–126 Koziol L, Schultz PA, House GL, Bauer JT, Middleton EL, Bever JD (2018) The
Ehrenfeld JG (2000) Defining the limits of restoration: the need for realistic goals. plant microbiome and native plant restoration: examples of native mycor-
Restoration Ecology 8:2–9 rhizal fungi. Bioscience 68:996–1006
Eloy L, Schmidt IB, Borges SL, Ferreira MC, dos Santos TA (2019) Seasonal fire Ladouceur E, Jiménez-Alfaro B, Marin M, Vitis MD, Abbandonato H,
management by traditional cattle ranchers prevents the spread of wildfire in Iannetta PPM, Bonomi C, Pritchard HW (2018) Native seed supply and
the Brazilian Cerrado. Ambio 48:890–899 the restoration species pool. Conservation Letters 11:e12381
Fagan ME (2020) A lesson unlearned? Underestimating tree cover in drylands Lamont BB, Wiens D (2003) Are seed set and speciation rates always low
biases global restoration maps. Global Change Biology 26:4679–4690 among species that resprout after fire, and why? Evolutionary Ecology
Fidelis A, Appezzato-da-Glória B, Pillar VD, Pfadenhauer J (2014) Does distur- 17:277–292
bance affect bud bank size and belowground structures diversity in Brazil- Le Stradic S (2012) Composition, phenology and restoration of campo rupestre
ian subtropical grasslands? Flora-Morphology, Distribution, Functional mountain grasslands - Brazil. PhD thesis. Universidade Federal de Minas Ger-
Ecology of Plants 209:110–116 ais & Avignon Université, https://hal.archives-ouvertes.fr/tel-00942593/
Fidelis A, Alvarado ST, Barradas ACS, Pivello VR (2018) The year 2017: mega- Le Stradic S, Buisson E, Fernandes GW (2015a) Baseline data for the conserva-
fires and management in the Cerrado. Fire 1:49 tion of Neotropical mountain grasslands: vegetation composition and struc-
Fornara DA, Du Toit JT (2008) Browsing-induced effects on leaf litter quality ture. Journal of Mountain Science 12:864–877
and decomposition in a Southern African savanna. Ecosystems 11:238–249 Le Stradic S, Silveira FAO, Buisson E, Cazelles K, Carvalho V, Fernandes GW
Franco M, Silvertown J (2004) A comparative demography of plants based upon (2015b) Diversity of germination strategies and seed dormancy in herba-
elasticities of vital rates. Ecology 85:531–538 ceous species of campo rupestre grasslands. Austral Ecology 40:537–546

16 of 18 Restoration Ecology April 2021

 Research agenda for tropical grassland restoration

Lugo MA, Pagano MC (2019) Overview of the mycorrhizal fungi in South Amer- Exclosure Experiment (KLEE). Pastoralism: Research, Policy and Practice
ica. Pages 1–27. In: Pagano MC, Lugo MA (eds) Mycorrhizal fungi in 2:10
South America. Springer International Publishing, Cham, Switzerland Riginos C, Porensky LM, Veblen KE, Young TP (2018) Herbivory and drought
Martins WA, Mantelli M, Santos SC, Netto APC, Pinto F (2015) Estaquia e con- generate short-term stochasticity and long-term stability in a savanna
centraç~ao de reguladores vegetais no enraizamento de Campomanesia ada- understory community. Ecological Applications 28:323–335
mantium. Revista de Ciências Agrárias 38:58–64 Rosbakh S, Pacini E, Nepi M, Poschlod P (2018) An unexplored side of regener-
Meli P, Holl KD, Benayas JMR, Jones HP, Jones PC, Montoya D, Mateos DM ation niche: seed quantity and quality are determined by the effect of tem-
(2017) A global review of past land use, climate, and active vs. passive res- perature on pollen performance. Frontiers in Plant Science 9:1036
toration effects on forest recovery. PLoS One 12:e0171368 Rusdy M (2017) A review on hardseedness and breaking dormancy in tropical
Midgley JJ, Lawes MJ, Chamaillé-Jammes S (2010) Savanna woody plant forage legumes. Livestock Research for Rural Development 29:237
dynamics: the role of fire and herbivory, separately and synergistically. Sampaio AB, Vieira DLM, Holl KD, Pellizzaro KF, Alves M, Coutinho AG,
Australian Journal of Botany 58:1–11 Cordeiro A, Ribeiro JF, Schmidt IB (2019) Lessons on direct seeding to
Miller BP, Sinclair EA, Menz MHM, Elliott CP, Bunn E, Commander LE, et al. restore Neotropical savanna. Ecological Engineering 138:148–154
(2017) A framework for the practical science necessary to restore sustain- Sankaran M, Hanan NP, Scholes RJ, Ratnam J, Augustine DJ, Cade BS, et al.
able, resilient, and biodiverse ecosystems. Restoration Ecology 25:605–617 (2005) Determinants of woody cover in African savannas. Nature 438:
Morecroft MD, Duffield S, Harley M, Pearce-Higgins JW, Stevens N, Watts O, 846–849
Whitaker J (2019) Measuring the success of climate change adaptation Schmidt IB, Eloy L (2020) Fire regime in the Brazilian savanna: recent changes,
and mitigation in terrestrial ecosystems. Science 366:9256 policy and management. Flora 268:151613
Nerlekar AN, Veldman JW (2020) High plant diversity and slow assembly of old- Schmidt IB, Ferreira MC, Sampaio AB, Walter BMT, Vieira DLM, Holl KD
growth grasslands. Proceedings of the National Academy of Sciences 117: (2019a) Tailoring restoration interventions to the grassland-savanna-forest
18550–18556 complex in Central Brazil. Restoration Ecology 27:942–948
Nevill PG, Tomlinson S, Elliott CP, Espeland EK, Dixon KW, Merritt DJ (2016) Schmidt IB, de Urzedo DI, Piña-Rodrigues FCM, Vieira DLM, de Rezende GM,
Seed production areas for the global restoration challenge. Ecology and Sampaio AB, Junqueira RGP (2019b) Community-based native seed pro-
Evolution 6:7490–7497 duction for restoration in Brazil – the role of science and policy. Plant Biol-
de Oliveira TG, Pina PSS, Bertoni BW, de França SC, Pereira AMS (2011) ogy 21:389–397
Micropropagaç~ao de Croton antisyphiliticus Mart. Ciência Rural 41: Seddon PJ, Griffiths CJ, Soorae PS, Armstrong DP (2014) Reversing defauna-
1712–1718 tion: restoring species in a changing world. Science 345:406–412
de Oliveira ACC, Forti VA, de Loiola PP, Viani RAG (2020) Techniques for Silva DB, Vieira RF, Cordeiro MCT, Pereira EBC, Pereira AV (2011)
seedling production of two native grasses: new perspectives for Brazilian Propagaç~ao vegetativa de Brosimum gaudichaudii Tréc. (mama-cadela)
Cerrado restoration. Restoration Ecology 28:297–303 por estacas de raízes. Revista Brasileira de Plantas Medicinais 13:151–156
Overbeck GE, Vélez-Martin E, Scarano FR, Lewinsohn TM, Fonseca CR, Silveira FAO, Arruda AJ, Bond W, Durigan G, Fidelis A, Kirkman K, et al.
Meyer ST, et al. (2015) Conservation in Brazil needs to include non-forest (2020) Myth-busting tropical grassy biome restoration. Restoration Ecol-
ecosystems. Diversity and Distributions 21:1455–1460 ogy, 28(5):1067–1073
Parr CL, Lehmann CER, Bond WJ, Hoffmann WA, Andersen AN (2014) Trop- Sitters J, Kimuyu DM, Young TP, Claeys P, Venterink HO (2020) Megaherbi-
ical grassy biomes: misunderstood, neglected, and under threat. Trends in vores reverse negative effects of cattle on soil carbon and nutrient pools.
Ecology & Evolution 29:205–213 Nature Sustainability 3:360–366
Parr CL, Eggleton P, Davies AB, Evans TA, Holdsworth S (2016) Suppression of Smith MD, Knapp AK, Collins SL, Burkepile DE, Kirkman KP, Koerner SE,
savanna ants alters invertebrate composition and influences key ecosystem et al. (2016) Shared drivers but divergent ecological dynamics: insights
processes. Ecology 97:1611–1617 from long-term experiments in savanna grasslands. Bioscience 66:666–682
Pausas JG, Lamont BB, Paula S, Appezzato-da-Glória B, Fidelis A (2018) de Souza AV, Oliveira FJV, Bertoni BW, França SC, Pereira AMS (2015) Enrai-
Unearthing belowground bud banks in fire-prone ecosystems. New Phytol- zamento in vitro de catuaba (Anemopaegma arvense (Vell.) Stell. ex de
ogist 217:1435–1448 Souza), uma planta medicinal do Cerrado. Revista Brasileira de Plantas
Perring MP, Standish RJ, Price JN, Craig MD, Erickson TE, Ruthrof KX, Medicinais 17:51–58
Whiteley AS, Valentine LE, Hobbs RJ (2015) Advances in restoration ecol- Staver AC, Bond WJ, Stock WD, van Rensburg SJ, Waldram MS (2009) Brows-
ogy: rising to the challenges of the coming decades. Ecosphere 6:art131 ing and fire interact to suppress tree density in an African savanna. Ecolog-
Pilon NAL, Buisson E, Durigan G (2018) Restoring Brazilian savanna ground ical Applications 19:1909–1919
layer vegetation by topsoil and hay transfer. Restoration Ecology 26:73–81 Stevens N, Lehmann CER, Murphy BP, Durigan G (2017) Savanna woody
Pilon NAL, Assis GB, Souza FM, Durigan G (2019) Native remnants can be encroachment is widespread across three continents. Global Change Biol-
sources of plants and topsoil to restore dry and wet cerrado grasslands. Res- ogy 23:235–244
toration Ecology 27:569–580 Suding KN, Gross KL, Houseman GR (2004) Alternative states and positive
Plante AF, McGill WB (2002) Soil aggregate dynamics and the retention of feedbacks in restoration ecology. Trends in Ecology & Evolution 19:46–53
organic matter in laboratory-incubated soil with differing simulated tillage Temperton VM, Buchmann N, Buisson E, Durigan G, Kazmierczak Ł,
frequencies. Soil and Tillage Research 66:79–92 Perring MP, de Dechoum MS, Veldman JW, Overbeck GE (2019) Step
Ploton P, Mortier F, Réjou-Méchain M, Barbier N, Picard N, Rossi V, et al. back from the forest and step up to the Bonn challenge: how a broad ecolog-
(2020) Spatial validation reveals poor predictive performance of large-scale ical perspective can promote successful landscape restoration. Restoration
ecological mapping models. Nature Communications 11:4540 Ecology 27:705–719
Prach K, Durigan G, Fennessy S, Overbeck GE, Torezan JM, Murphy SD (2019) du Toit JCO (2009) Early survival and growth of vegetatively propagated indig-
A primer on choosing goals and indicators to evaluate ecological restora- enous grasses in a clear-felled timber plantation in KwaZulu-Natal,
tion success. Restoration Ecology 27:917–923 South Africa. African Journal of Range and Forage Science 26:97–101
Pringle RM, Kimuyu DM, Sensenig RL, Palmer TM, Riginos C, Veblen KE, Tölgyesi C, Török P, Hábenczyus AA, Bátori Z, Valkó O, Deák B, Tóthmérész B,
Young TP (2015) Synergistic effects of fire and elephants on arboreal ani- Erd}os L, Kelemen A (2020) Underground deserts below fertility islands?
mals in an African savanna. Journal of Animal Ecology 84:1637–1645 Woody species desiccate lower soil layers in sandy drylands. Ecography
Riginos C, Porensky LM, Veblen KE, Odadi WO, Sensenig RL, Kimuyu D, 43:848–859
Keesing F, Wilkerson ML, Young TP (2012) Lessons on the relationship UN Environment Programme (2019) New UN decade on ecosystem restoration
between livestock husbandry and biodiversity from the Kenya Long-term offers unparalleled opportunity for job creation, food security and addressing

April 2021 Restoration Ecology 17 of 18

Research agenda for tropical grassland restoration

climate change. UNEP - UN Environ. Programme. http://www. Veldman JW, Silveira FAO, Fleischman FD, Ascarrunz NL, Durigan G (2017)
unenvironment.org/news-and-stories/press-release/new-un-decade-ecosystem- Grassy biomes: an inconvenient reality for large-scale forest restoration?
restoration-offers-unparalleled-opportunity (accessed 14 Oct 2019) A comment on the essay by Chazdon and Laestadius. American Journal
de Urzedo DI, Fisher R, Piña-Rodrigues FCM, Freire JM, Junqueira RGP (2019) of Botany 104:649–651
How policies constrain native seed supply for restoration in Brazil. Resto- Veldman JW, Aleman JC, Alvarado ST, Anderson TM, Archibald S, Bond WJ,
ration Ecology 27:768–774 et al. (2019) Comment on “the global tree restoration potential.” Science
Veblen KE (2012) Savanna glade hotspots: plant community development 366:eaay7976
and synergy with large herbivores. Journal of Arid Environments 78: Wagenius S, Beck J, Kiefer G (2020) Fire synchronizes flowering and boosts
119–127 reproduction in a widespread but declining prairie species. Proceedings of
Veblen KE, Porensky LM, Riginos C, Young TP (2016) Are cattle surrogate the National Academy of Sciences 117:3000–3005
wildlife? Savanna plant community composition explained by total herbiv- Waters C, Whalley W, Huxtable C (2001) Grassed up - guidelines for revege-
ory more than herbivore type. Ecological Applications 26:1610–1623 tating with Australian native grasses. NSW, Australia: NSWDPI, http://
Veldman JW (2016) Clarifying the confusion: old-growth savannahs and tropical www.dpi.nsw.gov.au/agriculture/pastures-and-rangelands/rangelands/
ecosystem degradation. Philosophical Transactions of the Royal Society of publications-and-information/grassedup (accessed 29 Sep 2020)
London. Series B, Biological Sciences 371:20150306 Werner CM, LaMalfa E, Sensenig RL, Kimuyu DM, Young TP (in press) Syner-
Veldman JW, Putz FE (2011) Grass-dominated vegetation, not species-diverse gistic effects of herbivores and repeat fire on spatial heterogeneity of pre-
natural savanna, replaces degraded tropical forests on the southern edge scribed burns and their consequences for tree saplings. Ecology
of the Amazon Basin. Biological Conservation 144:1419–1429 Young TP (2004) Fire-induced reproduction of Festuca pilgeri in the subalpine
Veldman JW, Buisson E, Durigan G, Fernandes GW, Le Stradic S, Mahy G, et al. zone of Mount Kenya. African Journal of Ecology 42:235–236
(2015) Toward an old-growth concept for grasslands, savannas, and wood- Zahawi RA, Reid JL, Holl KD (2015) Passive restoration can be an effective strat-
lands. Frontiers in Ecology and the Environment 13:154–162 egy: a reply to Prach and del Moral (2015). Restoration Ecology 23:347–348

Guest Coordinating Editor: Lars Brudvig Received: 27 May, 2020; First decision: 6 July, 2020; Revised: 16 September,
2020; Accepted: 22 September, 2020

18 of 18 Restoration Ecology April 2021

View publication stats