Trait-based approaches for guiding the restoration of ...
Transcript of Trait-based approaches for guiding the restoration of ...
Trait-based approaches for guiding the restoration of degraded agricultural landscapes in East Africa
Lohbeck, M., Winowiecki, L., Aynekulu, E., Okia, C., & Vågen, T. G.
This article is made publically available in the institutional repository of Wageningen University and Research, under article 25fa of the Dutch Copyright Act, also known
as the Amendment Taverne.
Article 25fa states that the author of a short scientific work funded either wholly or partially by Dutch public funds is entitled to make that work publicly available for no consideration following a reasonable period of time after the work was first published, provided that clear reference is made to the source of the first publication of the work.
For questions regarding the public availability of this article, please contact [email protected].
Please cite this publication as follows:
Lohbeck, M., Winowiecki, L., Aynekulu, E., Okia, C., & Vågen, T. G. (2018). Trait-based approaches for guiding the restoration of degraded agricultural landscapes in East Africa. Journal of Applied Ecology, 55(1), 59-68. DOI: 10.1111/1365-2664.13017
You can download the published version at:
https://doi.org/10.1111/1365-2664.13017
J Appl Ecol. 2018;55:59–68. wileyonlinelibrary.com/journal/jpe | 59© 2017 The Authors. Journal of Applied Ecology © 2017 British Ecological Society
Received:28November2016 | Accepted:13September2017DOI:10.1111/1365-2664.13017
F U N C T I O N A L T R A I T S I N A G R O E C O L O G Y
Trait- based approaches for guiding the restoration of degraded agricultural landscapes in East Africa
Madelon Lohbeck1,2 | Leigh Winowiecki1 | Ermias Aynekulu1 | Clement Okia3 | Tor-Gunnar Vågen1
1WorldAgroforestryCentre(ICRAF),Nairobi,Kenya2ForestEcologyandForestManagementGroup,WageningenUniversity,Wageningen,TheNetherlands3WorldAgroforestryCentre(ICRAF),Kampala,Uganda
CorrespondenceMadelonLohbeckEmail:[email protected]
Funding informationForests,TreesandAgroforestry;NederlandseOrganisatievoorWetenschappelijkOnderzoek,Grant/AwardNumber:863.15.017;IUCN,Grant/AwardNumber:IUCN-1124;CGIAR-FTA;AustralianCentreforInternationalAgriculturalResearch,Grant/AwardNumber:FSC/2012/014;InternationalFundforAgriculturalDevelopment,Grant/AwardNumber:2000000520
HandlingEditor:MarneyIsaac
Abstract1. Functionalecologyprovidesaframeworkthatcanlinkvegetationcharacteristicsofvarious land useswith ecosystem function. However, this application has beenmostly limitedto [semi-]naturalsystemsandsmall spatialscales.Here,weapplyfunctionalecologytofiveagricultural landscapesinKenya,UgandaandEthiopia,andasktowhatextentvegetationcharacteristicscontributetosoilfunctionsthatarekeytofarmers’livelihoods.
2. WeusedtheLandDegradationSurveillanceFramework(LDSF),amulti-scaleas-sessmentoflandhealth.EachLDSFsiteisa10×10kmlandscapeinwhichvegeta-tioncoveranderosionprevalenceweremeasured,atreeinventorywascarriedout,andtopsoil(0–20cm)sampleswerecollectedfororganiccarbon(SOC)analysisinapproximately160×1,000m2plots.Landdegradationisarecurringphenomenonacrossthefivelandscapes,indicatedbyhigherosionprevalence(67%–99%oftheplotswereseverelyeroded).Weusedmixedmodelstoassessifvegetationcover,above-groundwoodybiomassandthefunctionalpropertiesofwoodyvegetation(weighted-meantraitvalues,functionaldiversity[FD])explainvariationinSOCanderosionprevalence.
3. Wefoundthatthevegetationcoverandabove-groundbiomasshadstrongpositiveeffectsonsoilhealthbyincreasingSOCandreducingsoilerosion.Aftercontrollingforcoverandbiomass,wefoundadditionalmarginaleffectsoffunctionalproper-tieswhereFDwaspositivelyassociatedwithSOCandtheabundanceofinvasivespecieswasassociatedwithhighersoilerosion.
4. Synthesis and applications.Thisworkillustrateshowfunctionalecologycanprovidemuch-needed evidence for designing strategies to restore degraded agriculturallandandtheecosystemservicesonwhichfarmersdepend.Weshowthattoensuresoilhealth,itisvitaltoavoidexposedsoil,maintainorpromotetreecover,whileensuringfunctionaldiversityoftreespecies,andtoeradicateinvasivespecies.
K E Y W O R D S
agriculturalland,agroecology,agroforestry,erosion,functionaldiversity,functionaltraits,landdegradation,soilhealth,soilorganiccarbon,vegetation
60 | Journal of Applied Ecology LOHBECK Et aL.
1 | INTRODUCTION
Thenegativeimpactsoflanddegradationonproductivity,biodiversityand local livelihoods have become undeniable (Pereira etal., 2010;Pimentel&Burgess,2013).Asaconsequence,restoration,heredefinedas the practice of assisting the recovery of degraded ecosystems, isnowaglobalpriority (Minnemeyer,Laestadius,Sizer,SaintLaurent,&Potapov,2011).Restorationprovidesopportunitiestocounteractdeg-radationandreviveecosystemfunctions,includingcomponentsofbio-diversity(Benayas,Newton,Diaz,&Bullock,2009;Chazdon,2008)andsoilfertility,whichiskeytofarmers’livelihoods(Diemontetal.,2006).
In this study, we assess degradation in agricultural landscapesusingtwomainindicators:soilorganiccarbon(SOC)anderosionprev-alence.SOC isawidelyused indicatorofsoilhealthas it influencesseveralimportantsoilpropertiessuchascationexchangecapacityandwaterholdingcapacity(Lal,Griffin,Apt,Lave,&Morgan,2004).Soilerosion is an indicatorof landdegradation and is includedas a keyprocessleadingtolossofSOCanddecliningsoilhealthandproduc-tivity(Dregne,2002).Bothindicatorsareheavilyinfluencedbyman-agement,andunsustainablelandusehasbeenshowntoreduceSOCand increaseerosion,making these suitable indicators for assessinglanddegradationandsoilhealth(Dregne,2002;Laletal.,2004;Vågen,Winowiecki,Abegaz,&Hadgu,2013;Winowieckietal.,2015).
Increasingtreecoverisacoreactivityforrestoringdegradedlands(Lamb,Erskine,&Parrotta,2005).Recentevidenceshowsthatincreas-ingtreecover inthedrytropicscan improvesoil function, includingwateravailability(Ilstedtetal.,2016).Furthermore,increasingwoodybiomasspositivelyaffectsproductivityandlitterdecompositionratesinregeneratingforests(Lohbeck,Poorter,Martínez-Ramos,&Bongers,2015)andSOCinagroforestrysystems(Hombegowda,vanStraaten,Köhler,&Hölscher,2016;Lorenz&Lal,2014).However,theinfluenceoftreesonsoilhealthmaydifferfordifferenttreespecies,andunder-standing this is crucial fordesigningeffective restoration strategies.Insightscanbegainedfromthefieldoffunctionalecology(Laughlin,2014;Sandel,Corbin,&Krupa,2011),whichprovidesaframeworktomechanistically link landusewithspecies’ functional traitsandeco-systemfunction(e.g.Cadotte,Carscadden,&Mirotchnick,2011;Díazetal.,2007;Lavoreletal.,2010).
Plantfunctionaltraits,andatacoarsebiologicalscalefunctionaltypes,areindicatorsofplantstrategiesandofhowspeciesinfluenceecosystem function (Petchey & Gaston, 2006). Accordingly, many
plant functional traits and types contribute to soil health (Table1).Wooddensity, for instance, indicates species’ positioningalong the“resource-economics spectrum” (Chave etal., 2009). High-wooddensityspecieshaveexpensive-to-constructtissuesthatdecomposeslowly,andtherebyhaveamoreconstantandlastingpositiveeffectonSOCinputs(deDeyn,Cornelissen,&Bardgett,2008).Functionaltraitsthatdescribethearchitectureoftreesmayinfluencesoilhealthby altered understorey climatic conditions. For instance, trees thathaveatallandnarrowgrowthformwillshadethesoiltoalesserex-tentandmayincreasetemperature,decreasesoilmoistureandnega-tivelyaffectsoilhealth(Chapin,2003;Linetal.,2016).Furthermore,certain functional types are known to have specific effects on soilhealth.Treesabletofixatmosphericdinitrogen(N2)dosobymutu-alisticsymbiosiswithbacteria,resultinginfastergrowth(Battermanetal.,2013)andenhancedsoilhealth(e.g.Adams,Turnbull,Sprent,&Buchmann,2016;Bradfordetal.,2002).Deciduousspeciesundergoleaf senescence forpartof theyear, therebyproducing largequan-titiesof litterfororganic-carbon inputs intothesoil (deDeynetal.,2008).Incontrast,somefunctionaltypesareknownfortheirnegativeimpactsonsoilhealth:invasivespecieshavebeenassociatedwithin-creasederosion(Grover&Musick,1990;Vågen&Winowiecki,2014),decreasedecosystemcarbon(Jackson,Banner,Jobbágy,Pockman,&Wall,2002)anddecreasedstreamflow(Cleverly,Smith,Sala,&Devitt,1997).Also commonlyplanted exotics such asEucalyptus spp.mayreduceunderstoreyvegetationcoveranddiversity(Thijsetal.,2014)andnegativelyimpacthydrology(Zhou,Morris,Yan,Yu,&Peng,2002;butseeReynolds,Wassie,Wubalem,Liang,&Collins,2016).
Besides predictions on how species-level functional traits andtypes influenceecosystem function, twomain theoriesexplainhowthe traits of species co-occurring in a community (community-levelfunctional properties) influence ecosystem function.Themass-ratiohypothesispredictsthatthetraitsofthedominantspeciesdrivefunc-tions(Grime,1998),whilethenichecomplementarityhypothesispre-dicts that functionally diverse communities are better able tomakeoptimaluseofavailableresourcesandtherebyincreaseoverallfunc-tionality(e.g.Cardinaleetal.,2012).
We evaluate the extent towhich vegetation contributes to soilhealth.We do so by assessing a hierarchy of vegetation indicatorsthatreflectincreasinglydetailedcharacteristicsofthevegetationandthereby systematically assesswhat aspects ofvegetation should bepromotedforrestoringdegradedlandscapes.
TABLE 1 Summaryofthehypothesizedrelationshipsbetweenfunctionaltraits/typesandsoilhealth.+/−indicatepositive/negativepredictedeffectsonsoilhealth,indicatedbySOC(soilorganiccarbon)(positively)anderosion(negatively)
Functional trait/ type Plant strategies and ecosystem function
Effect on soil health
Wooddensity Conservativestrategy,slowgrowth,slowdecomposition,above-groundbiomass +
Adultheight Lightdemanding,moreevapotranspiration,above-groundbiomass,tallarchitecturecausinglessshading −
N2-fixing Fastgrowth,highfoliarnitrogen,N-mineralization,soilnitrification +
Deciduous Lessevapotranspiration,fasterdecomposition,morelitterproduction,shallowroots,highwooddensity +
Invasive Out-competingoriginalvegetationcover,fastgrowthandreproduction −
Exotic Fastgrowth,lightdemanding,reducedsoilwateravailability −
| 61Journal of Applied EcologyLOHBECK Et aL.
We hypothesize that: (i) increased vegetation cover reducessoildegradation(increasesSOCanddecreaseserosion); (ii)above-groundwoodybiomassreducessoildegradation;and(iii)functionalpropertiesofthevegetationaffectsoildegradation.Specifically,(a)increasedfunctionaldiversity(FD)reducessoildegradation,(b)par-ticularfunctionaltraits(highwooddensity,lowadultheight)reducesoildegradation,and(c)particularfunctionaltypesofwoodyvege-tation (N2-fixers,deciduousspecies) reducesoildegradationwhileotherfunctionaltypes(invasivespecies,exoticspecies)increasesoildegradation.
2 | MATERIALS AND METHODS
2.1 | Study sites
Thestudytookplaceinfiveagriculturallandscapesinthreecountriesin East Africa (Figure1). All landscapes are characterized by small-holder farming systems and are degraded, indicated bywidespreaderosion. Table S1 summarizes key climatic variables and vegetationtypesperlandscape,whileFigureS1givesthevariationinvegetationstructurefoundacrosslandscapes.
In Uganda, we focused on two landscapes in eastern Uganda,bordering Mount Elgon National Park: Mbale (34.24E, 1.09N) andBumagabula (34.39E, 1.16N).The area is characterized by amoun-tainoustopography,whereBumagabulaislocatedathigherelevationandhashigherrainfallthanMbale.Maize,legumes,bananaandcoffeeare commonly cultivated, often in agroforestry systems,with someeucalypt plantations and cattle grazing areas. The region has highpopulationdensities,estimatedat620personsperkm2in2002(UBS,
2012).InEthiopia,wefocusedontwolandscapes,thesubhumidAno(36.97E,9.09N)andthesemi-aridAlemTena(38.90E,8.24N).Inbothsites,themaincropsweresorghum,maizeandteff,withtreescom-monlyintegratedintofarmingsystems(Iiyamaetal.,2016).InKenya,wefocusedononelandscape,Waita(38.19E,0.91S),inKituicounty.Thisisalowlandsitewheresmallholderfarmerscultivatemaize,milletandsorghumwithsmall-scalecattleproduction.Waitaisthedriestofourlandscapeswithanannualrainfallof767mmperyear.
2.2 | Sampling framework
TheLandDegradationSurveillanceFramework(LDSF)wasusedtoas-sessbiophysicalindicatorsatthefivelandscapesites.TheLDSFusesahierarchicalsamplingframework;eachsiteis100km2,andconsistsofsixteen1-km2clusters,eachclusterconsistsoften1,000-m2sam-pling plots and each plot consists of four 100-m2 subplots (Vågen,Winowiecki, Tamene Desta, & Tondoh, 2013). Positioning of siteswasbasedonongoingprojectactivitiesinareasofinterest.Locationswererandomizedtocovervariationintopographyandlanduseswhileavoiding lakes and rivers. The LDSF is designed for simultaneouslyassessingkey indicatorsofecosystemhealthacrossmultiple spatialscalesandatgeo-referencedlocations.
2.3 | Soil health indicators
Soilerosionprevalencewasscoredateachsubplot(n=640observa-tionspersite),whenerosionwasobservedinoverhalfofthefoursub-plotsperplot,thisplotwasconsideredtobeseverelyeroded(binary0/1).Topsoilsamples (0–20cm)werecollectedateachsubplotand
F IGURE 1 MapsofthefivestudylocationsacrossthreecountriesinEastAfrica.SeeTableS1andmethodsformoreinformation
MbaleBumagabula
Uganda
Alem Tena
Ano
Ethiopia
Waita
Kenya
62 | Journal of Applied Ecology LOHBECK Et aL.
thoroughlymixedtoformacompositetopsoilsampleforeachplot.SOCandsandcontentweremeasuredthroughMIRabsorbance,de-tailedmethodsofwhicharepresentedinAppendixS1.Mid-infraredspectroscopy is becoming awell-establishedmethod for predictingsoil properties (cf.Madari etal., 2006; Reeves, Follett,McCarty, &Kimble,2006;Vågen,Winowiecki,Abegaz,etal.,2013).Tenpercentofthesoilsamplescollectedateachsitewereconsideredreferencesamples(n=32persite)andwereanalysedforSOCandsandcontent.Calibrationmodelsweredevelopedforthepredictionofsoilproper-tiesusingMIRspectra from the ICRAFpan-AfricanMIRspectral li-braryandtheresultsofsoilanalysisonthereferencesamples(Vågen,Winowiecki,Abegaz,etal.,2013;Vågen,Winowiecki,Tondoh,Desta,&Gumbricht,2016).Thismethodhasbeenshowntoaccuratelypre-dictSOCacrossSub-SaharanAfrica(Vågenetal.,2016).
2.4 | Vegetation cover and biomass estimations
Vegetationcoveringthesoil(mainlyherbsandgrasses)wasratedineachofthesubplotsusingaBraun–Blanquetvegetationratingscalethatrangesfrom0(exposedsoil)to5(>65%cover;Braun-Blanquet,1932).Plot-levelvegetationcoverrepresentsthemeanofthevegeta-tioncoverclassesfromthefoursubplots.Treeinventorieswerecar-riedoutinslightlydifferentwaysdependingonthesite,asexplainedindetailinAppendixS2.Weestimatedplot-levelabove-groundbiomass(Mg/ha)usingagenericallometricformulabasedonthediameteratbreastheight(DBH),species-specificwooddensityandasite-specific“environmentalstressfactor”(Chaveetal.,2014).Thiswasexpressedonaper-hectarebasisasandisthuscorrectedfordifferencesinplot-levelsamplingeffortacrossthesitesandplots.
2.5 | Functional properties of the woody vegetation
Atotalof2,673treesbelongingto137differentspecieswereidenti-fiedacrossthefivelandscapes.Dataforanumberofrelevantfunc-tionaltraitsandtypeswereretrievedfromflorasandonlinesourcesforthetreespecies:Wooddensity(g/cm3),adultheight(m),N2-fixing(0/1),deciduous(0/1), invasive(0/1)andexotic (0/1),forwhichde-tailedmethodsarepresentedinAppendixS3.
Species-levelfunctionaltraitswerescaledtoplot-levelfunctionalproperties using two complementary metrics: community-weightedmean (CWM)andFD.TheCWM (Garnier etal., 2004) is calculatedbasedoneachsingle traitor typeandweightedbyspecies’ relativebasalareaintheplot.Forcontinuoustraitvalues,theCWMreflectsthe trait value of “theweighted-averagewoody plant” in the com-munity, forbinaryvariables this reflects theproportionof thebasalareathat isrepresentedbythattype.FDwascalculatedusingRao’squadraticentropy(Rao’sQ) (Botta-Dukát,2005)andisbasedonthefunctionaldistancebetweenspeciesweightedbytheirrelativebasalareas,makinguseofalltraitssimultaneously.Rao’sQisconceptuallysimilartofunctionaldispersion(Laliberté&Legendre,2010)andesti-mateshow functionallydifferent theco-occurring speciesare.Plot-levelfunctionalpropertieswerecalculatedusingtherpackage“FD”(Laliberté&Shipley,2012).
2.6 | Statistical analysis
In this study,we took theplot as aunit of replication,with a totalof745plots.Weusedgeneralized linearmixedmodels, from ther package“lme4”(Bates,Maechler,Bolker,&Walker,2015)tosystem-aticallytestfortheeffectsofvegetationonsoilhealthinaseriesofmodelsthatreflectincreasedcomplexity(Table2).
Mixed-effectsmodelsenableaccountingfordifferencesincross-sitesamplingdesign,bytakingsiteasarandomeffect,allowingarandomintercept for each site. With package “LMERConvenienceFunctions”(Tremblay&Ransijn,2015),weconfirmedthatsiteindeedcontributedasarandomeffect.Inmodel5,wesystematicallyreplacedthedifferentplot-levelfunctionalproperties(6CWM+1FD=7variationsonmodel5),resultingin12modelspersoilhealthindicatorand24modelsintotal.
Themodelwith thebest fitwas selectedbasedonAkaike infor-mation criterion, adjusted for small sampling size (AICc) (Burnham&Anderson,2002).AICpenalizesformodelcomplexity,hencetakingaconservativeapproachtoassessingtheimpactsoftreesandfunctionaltraitsonsoilhealth.Whenmodelsdidnotdiffersignificantly(ΔAICc < 2),wechosethemodelthathadthehighestmarginalandconditionalR2 (Nakagawa,Schielzeth,&O’Hara,2013), computedusingpackage“piecewiseSEM”(Lefcheck,2015).Forsevereerosion(binary,0/1),weusedglmer(family=binomial)whileforSOC(continuous,range3–96g/kg)weusedlmer.Modelstatisticswerederivedusingpackages“sjstats”and “sjPlot” (Lüdecke,2016a,2016b),while significance levels reflectthez-associatedp-value (forerosion),or the t-associatedp-value (forSOC)derivedusing“nlme” (Pinheiro,Bates,DebRoy,&Sarkar,2016).Allanalyseswerecarriedoutusingrversion3.2.4(RCoreTeam,2014).
3 | RESULTS
3.1 | Site conditions
ThefiveEastAfricanstudysitesrepresentalargevarietyofclimatic,topographicalandland-usecharacteristics(FiguresS1andS2).Erosionwaswidespreadacrossthesites(67%–99%acrosseachlandscape),in-dicatingtheneedformoresustainablelandmanagementpracticesandland restorationactivities.Average topsoilOCwas29.8g/kg±13.2forBumagabula,27.9g/kg±4.2 forAno,21.2g/kg±8.3 forMbale,14.3g/kg±4.0forAlemand10.1g/kg±4.0forWaita(FigureS2).
3.2 | Optimal model
The most complex model, with the largest number of variables(Table2,model5),bestexplainedSOCandsoilerosion.Thismodelincludedsoiltexture(sandcontent),vegetationcover,above-groundwoodybiomassand functionalpropertiesof thewoodyvegetation.Wefoundthatsoilhealth(lowererosionandhigherSOC)wasassoci-atedwithhighervegetationcoverandhigherabove-groundbiomass,asexpected.Aftercontrollingforthese,wefoundthatdistinctfunc-tional properties related to distinct aspects of soil health; invasivespecieswereassociatedwith increasederosionwhileFDwasasso-ciatedwith increasedSOC (Figures2and3,Table3).Althoughour
| 63Journal of Applied EcologyLOHBECK Et aL.
modelselectionsuggestsaroleforfunctionalpropertiesofthewoodyvegetationinexplainingsoilhealth,theirmarginaleffectsalonewerenotsignificant.Thevarianceexplainedbythetotalmodelforsevereerosionwas40%(32%forfixedfactorsalone),whilethevarianceex-plainedforSOCwas56%(11%forfixedfactorsalone).Modelfitdidnotimprovewhenallowingthesitestodifferinthevegetationindica-tors’ fixedfactoreffects,suggestingthattheeffectsfoundarecon-sistentacrossthesites.TableS2givestheinterceptsacrossthesites.
4 | DISCUSSION
Restoration of agricultural landscapes provides an opportunity to in-creasetheproductivityandresilienceofagriculturalsystemsandsimul-taneouslycontributetoconservationobjectives.Functionalecologyis
a promising tool to guide science-based restoration (Laughlin, 2014)though its application to managed agricultural landscapes has beenlagging(Woodetal.,2015).Inthisstudy,weappliedatrait-basedap-proachtosoilhealthindegradedagriculturallandscapesandfoundthatthemarginaleffectsof thevegetationandtheir functionalpropertiesweredirectionallyintuitiveandhadclearimplicationsforrestoration.
4.1 | Vegetation effects on soil health
Wefound thatvegetationcoverandabove-groundbiomassare im-portantforsoilhealthashighervalueswereassociatedwithincreasedSOCanddecreasederosion.Wealsofoundmarginaladditionaleffectsforthefunctionalpropertiesofthewoodyvegetation.Invasivespecieswereassociatedwithincreasederosion,whileFDwasassociatedwithincreasedSOC.
TABLE 2 Themodelstestedinthisstudythatreflectincreasinglydetailedinformationonthevegetationtoexplainsoilhealth(erosionandSOC(soilorganiccarbon)).Givenaretherationaleforeachmodelandtheimplicationsforrestoration
# Model Rationale Implications for restoration
1 Soilhealth~Intercept Datacannotexplainsoilhealth None
2 Soilhealth~Sandcontent Soiltextureexplainssoilhealth None
3 Soilhealth~Sandcontent+Vegetationcover Vegetationcovercontributestosoilhealth Promotevegetationcover
4 Soilhealth~Sandcontent+Vegetationcover+Above-groundwoodybiomass
Above-groundbiomasscontributestosoilhealth
Plantandpromotetrees
5 Soilhealth~Sandcontent+Vegetationcover+Above-groundwoodybiomass+Functionalproperties(CWM/FD)
Functionalpropertiescontributetosoilhealth See5aand5b
5a Soilhealth~Sandcontent+Vegetationcover+Above-groundwoodybiomass+Community-weightedmeanfunctional-traitvaluesa
Functionaltraitsofthedominantspeciescontributetosoilhealth(mass-ratioeffect)
Plantandpromotespecificfunctionaltypesoftrees(andavoidothers)
5b Soilhealth~Sandcontent+Vegetationcover+Above-groundwoodybiomass+Functional-traitdiversity
Functionaldiversitycontributestosoilhealth(nichecomplementarityeffect)
Plantandpromoteadiverserangeoffunctionaltypesoftrees
aCWMsarecalculatedforsingletraits,sothismodelwastestedforeachofthesixfunctionaltraitsandtypes,seeTable1forspecifichypotheses.
F IGURE 2 Marginaleffectsoffixedeffectspredictingtheprobabilityofencounteringsevereerosion
0 10 20 30 40 50 60 70
0.0
0.2
0.4
0.6
0.8
1.0
Sand content (%)
Pre
d. p
rob.
of s
ever
e er
osio
n
0 1 2 3 4 5
0.0
0.2
0.4
0.6
0.8
1.0
Vegetation cover score
Pre
d. p
rob.
of s
ever
e er
osio
n
0 1 2 3 4 5 6
0.0
0.2
0.4
0.6
0.8
1.0
Above-ground biomass (Mg/ha)
Pre
d. p
rob.
of s
ever
e er
osio
n
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Rel. basal area invasive spp.
Pre
d. p
rob.
of s
ever
e er
osio
n
64 | Journal of Applied Ecology LOHBECK Et aL.
Our results substantiated that functional traits affect soil carbon(deDeynetal.,2008)anderosion (Lorenz&Lal,2005;Stokes,Atger,Bengough,Fourcaud,&Sidle,2009).Ourfindingssuggestthatthemech-anismbywhichthefunctionalpropertiesinfluencesoilhealthdependsontheindicator;wefoundthaterosionresistanceisdrivenbythetraitsofthedominantspecies(mass-ratioeffect),whileSOCwasdrivenbythediversityoftraitsintheecosystem(nichecomplementarityeffect).
4.1.1 | Erosion
Above-groundvegetationquantity(coverandbiomass)isdirectlyre-latedtobelow-groundvegetationquantityand,notsurprisingly,root
quantityanddistributioninthesoilareofhugeimportancetopreventerosion (e.g.DuránZuazo&RodríguezPleguezuelo, 2008;Gyssels,Poesen,Bochet,&Li,2005;Stokesetal.,2009).Therearelargeinter-specificdifferencesineffectsonsoilstability(Berendse,vanRuijven,Jongejans,&Keesstra,2015;Stokesetal.,2009),whichmaybedrivenbydifferences inspeciestraits.Wefoundthathigherabundanceofinvasive species was associated with increased erosion, suggestingthatthetraitsofthedominantspecies,andnotthediversity,explainerosion. Increased erosion under invasive species has been repeat-edly documented (Grover & Musick, 1990; Kourtev, Ehrenfeld, &Häggblom,2002;Vågen&Winowiecki,2014).Possiblemechanismsincludethatinvasivespeciestendtoinvestlessinsoil-stabilizingroot
F IGURE 3 Marginaleffectsoffixedeffectspredictingsoilorganiccarbon
0 10 20 30 40 50 60 70
010
2030
40
Sand content (%)
Soi
l org
anic
car
bon
(g/k
g)
0 1 2 3 4 5
010
2030
40
Vegetation cover score
Soi
l org
anic
car
bon
(g/k
g)
0 1 2 3 4 5 6
010
2030
40
Above-ground biomass (Mg/ha)
Soi
l org
anic
car
bon
(g/k
g)
0.00 0.02 0.04 0.06 0.08 0.10 0.12
010
2030
40
Functional diversity (Rao's Q)
Soi
l org
anic
car
bon
(g/k
g)
TABLE 3 Fixed-effectsstatisticsfortheoptimalmodelsexplainingsoilhealth:(a)severeerosionprevalenceand(b)SOC(soilorganiccarbon).Givenarethebetaestimates,theoddsratioandassociatedconfidenceintervals(forerosion)orstandardizedbetaestimateandassociatedconfidenceintervals(forSOC).p-valuesreflectthez-associatedp-value(forerosion),orthet-associatedp-value(forSOC).Site(#=5)wasincludedasarandomeffectforallmodels,totalN=745
(a) Severe erosion (R2conditional
0.40, R2marginal
0.32)
Predictor Estimate Odds ratio CI p
Intercept 3.73 41.59 15.79to109.57 <.001
Sandcontent 0.005 1.01 0.99to1.02 .546
Vegetationcover −0.708 0.49 0.39to0.62 <.001
Above-groundbiomass −0.536 0.59 0.16to2.19 .427
CWMinvasives 0.919 2.51a 0.62to10.13 .197
(b) Soil organic carbon (SOC) (R2conditional
0.56, R2marginal
0.11)
Predictor Estimate CI Std. estimate CI p
Intercept 22.4 16.41to28.43 <.001
Sandcontent −0.28 −0.33to−0.23 −0.35 −0.42to−0.29 <.001
Vegetationcover 0.89 0.32to1.47 0.17 0.06to0.28 .014
Above-groundbiomass 3.89 0.22to7.57 0.05 0.00to0.10 .038
Rao’sQ 4.12 −18.97to27.20 0.01 −0.04to0.06 .726
aProbabilityoferosionunderinvasivespeciesisthen(41.59×2.51)/(1+41.59×2.51)=0.99.
| 65Journal of Applied EcologyLOHBECK Et aL.
biomasscomparedwithnoninvasivespecies(vanKleunen,Weber,&Fischer,2010)andthat invasivespecies inhibitunderstoreyvegeta-tioncover.Althoughtheeffectof invasivespecieswasby itselfnotstatisticallysignificant,theeffectssizesuggestedthataninvadedsitehasa99%chancetobeseverelyeroded(Table3).Thisisnoteworthygiventhatinvasivepropertiesofaspecies,asfoundintheliterature,reflectthespecies’potentialtoinvadeandnotwhetheritisactuallyinvadingthesite.Besidesbeingapotentialdriverofdegradation,in-vasivespeciescanalsobeasymptomofdegradation.Possibleposi-tivefeedbackmechanismsregardinginvasivespeciesanddegradationcouldpotentiallyleadtoirreversibledegradationifrestorationeffortsare not implemented in time. Our result confirms that decreasingtheabundanceofinvasivespeciesshouldbeapriorityinrestorationefforts.
4.1.2 | Soil organic carbon
Soilcarbonstocksresultfromthebalancebetweencarboninputviaprimary productivity and carbon output via decomposition, volatili-zation (e.g. by charring or burning), leaching and erosion of topsoil(Amundson,2001).Wefoundthatvegetationcoverandbiomassin-creasedSOC.Indeed,coverandbiomassreduceerosion,asdiscussedintheprevioussection.Above-groundbiomassisadriverofprimaryproductivity (Lohbeck etal., 2015), although it may also acceleratedecomposition by enhancing soilmoisture by reducing evaporation(Lebrija-Trejos, Pérez-García, Meave, Poorter, & Bongers, 2011).Further,morebiomassgenerallyproducesmorelitter(Lohbecketal.,2015), providingaprimary input forSOC.Wealso foundaneffectof FD on SOC, suggesting that resource-use complementarity in aplantcommunity,possiblyincombinationwithfacilitation,enhancesSOCcontent.Previousresearchsimilarlyreportedthenichecomple-mentarityeffecttobeamajordriverofSOCinexperimentalgrass-lands (Fornara&Tilman,2008)and inagroforestrysystems in India(Hombegowda etal., 2016). In contrast, a recent study in Chinesesubtropical forest showed that SOC was mainly influenced by thecommunity-weightedmaximumheight of the trees, and less byFD(Linetal.,2016).
Consistentwithfunctionalecologytheory(Díazetal.,2007),ourresults suggest that functional traitsplay a role in carbondynamicsbymediatingspeciesdifferences inproductivityanddecomposition.Empirical evidence supports that niche complementarity drives pri-maryproductivityintropicalforest(Haggar&Ewel,1997)aswellasin temperate grasslands (Wilsey & Potvin, 2000). Other studies in-steadsupportthemass-ratiohypothesisshowingthatthefunctionaltraitsofthedominantspeciesdriveproductivity(Paquette&Messier,2011;Warren,Topping,&James,2009).Similarlyforlitterdecompo-sition,studieshavefoundbothdiversityeffects(Finertyetal.,2016;Scherer-Lorenzen,2008)andeffectsofthetraitsofdominantspeciesondecompositionrates(Garnieretal.,2004;Tardif&Shipley,2013).Probably both mechanisms matter for ecosystem function (Handaetal.,2014;Lohbecketal.,2015).Ourdiversity-effectcouldindicatea direct diversity-effect of vegetation on SOC through productivityanddecomposition (Hooperetal.,2005),butalsoan indirecteffect
mediatedbysoilbiota(Zak,Holmes,White,Peacock,&Tilman,2003).Thissuggeststhatwhenfarmersdecidetoplanttreesontheirfields,itisbeneficialtochoosespeciesthatarefunctionallycomplementarytotheonesalreadyestablished.
4.2 | Small marginal effects of functional properties
The variances explainedby the fixed effectswere quite small, par-ticularlyforerosion(Rm0.11).Highlevelsofsevereerosionacrossourlandscapes (67%–99%) reduced the variation in which vegetation-effectscouldbedetected.Ouralternativemodelsweredesignedtoreflect increasinglydetailedaspectsofthevegetation,takingacon-servative approach to themarginal effects of functional properties,whichpartlyexplainswhyeffectsweresmallandstatisticallynotsig-nificant(Table3).Itisimportanttorecognizethatthisobservationalstudyrepresentsalargevariationof landscapesshapedbydifferentpeopleandlandmanagementpractices.Thereisagreatneedtotestwhetherfunctional-traiteffectsonsoilfunctionscanbedetectedindynamichuman-modified landscapes,andwhat the implicationsareforrestoration,whichiswhatweexploredinthisstudy.Althoughthemarginaleffectsof functionalpropertiesare small,weconsiderourfindingsimportantbecausefunctionalpropertiesofthevegetationcaneasilybemodifiedbyselectingspecieswithsuitablefunctionaltraitswhenplantingtreesonfarmland.Thisapproach,thus,contributestoamuch-neededevidence-baseforrestoringagriculturallandscapes.
4.3 | Synthesis and applications
Based on our findings,we are able to draw recommendations thatwill advance the fieldof functional ecology inmanagedagriculturallandscapes.We showed that (nonwoody) vegetation cover stronglyinfluencedsoilproperties,suggestingthat includingfunctional traitsof nonwoody vegetation will increase our understanding of trait-mediated effects of vegetation on soils. Besides the direct effectsthatplantsexertonsoilfunctions,therearesomeimportantindirectlinkages between plants and the soil, mediated through manage-ment, symbionts and soil biota.Managementpractices, such as till-age,theuseoffireandfertilizers,werenotincludedinouranalyses.Managementdirectlyaffectssoilfunctionbutalsoindirectlythroughthe vegetation. We were constrained to functional traits availablefromonline sources and floras,which is a limited subset of above-ground traits and limited towoodyvegetation.Below-groundplanttraits(relatedtorootbiomassandturnover)areofparticularimpor-tance for soil functions (McCormack etal., 2015; Prieto, Stokes, &Roumet,2016;Schroth,1995).Futureresearchonfunctionalecologyinagriculturallandscapeswillneedtoincludetraitsofnonwoodyandcultivatedspecies,andmoreexplicitlyincludethedirectandindirecteffectsofmanagementonplantcommunitiesandonsoilhealth.
Understanding the functionalecologyofmanagedsystems isanimportant step towards making informed decisions on restorationplanning,bothattheplot-levelandatlandscape-scale.ApplyingthisapproachtodegradedEastAfricanlandscapes,wesuggestthatinad-ditiontoavoidingexposedsoilandpromotingtreesonfarms,priority
66 | Journal of Applied Ecology LOHBECK Et aL.
shouldbegiventotheremovalofinvasivespeciesandpromotionofhigherFDoftreesonfarmsforrestoringimportantsoilfunctionssuchasSOCandincreasedresistancetoerosion.
ACKNOWLEDGEMENTS
Wearegratefultothecommunitiesandlocalofficialsfortheirsup-portandforgrantingpermissiontocollectdataontheirproperties.AcknowledgementsareextendedtoJohnThiongoMaina,HumphreyWanjohi, Kabonesa Bernadette, Sam Chemusto and TesmesgenYohannesduringfieldcampaigns,aswellastothestaffoftheICRAFSoilandPlantSpectroscopyLaboratory inNairobi forprocessingofsoilsamples.WethankFaithMusiliforassistancewiththetraitdata-baseandDanielLüdeckeandEdwinLebrija-Trejosforstatisticalad-vice.FieldworkandsoilanalysiswerefundedbyIUCN(IUCN-1124),ACIAR (Trees for food security; FSC/2012/014), IFAD (Restorationof Degradation Lands for Food Security and Poverty Reduction inEastAfricaandtheSahel;2000000520)andtheCGIARProgramonForests, Trees and Agroforestry (FTA). M.L. was supported by re-searchprogrammeALW (863.15.017), financedby theNetherlandsOrganisationforScientificResearch(NWO).
AUTHORS’ CONTRIBUTIONS
M.L.,T.-G.V.andL.W.conceived the ideasanddesigned themeth-odology.M.L., T.-G.V., L.W., E.A. andC.O. collected thedata.M.L.,T.-G.V.,L.W.analysedthedata,M.L.ledthewritingofthemanuscript.Allauthorscontributedcriticallytothedraftsandgavefinalapprovalforpublication.
DATA ACCESSIBILITY
LDSF sites data are available from Harvard Dataverse: Ethiopia:https://doi.org/10.7910/dvn/29652 (Aynekulu & Shepherd, 2015),Kenya: https://doi.org/10.7910/dvn/sbl27o (Winowiecki & Sinclair,2016), Uganda: https://doi.org/10.7910/dvn/eqfm2n (Vågen etal.,2017).
ORCID
Madelon Lohbeck http://orcid.org/0000-0002-3959-1800
Leigh Winowiecki http://orcid.org/0000-0001-5572-1284
Ermias Aynekulu http://orcid.org/0000-0002-1955-6995
Tor-Gunnar Vågen http://orcid.org/0000-0002-8699-731X
REFERENCES
Adams, M. A., Turnbull, T. L., Sprent, J. I., & Buchmann, N. (2016).Legumes are different: Leaf nitrogen, photosynthesis, and wateruseefficiency.Proceedings of the National Academy of Sciences of the United States of America,113,4098–4103.
Amundson,R. (2001).Thecarbonbudget insoils.Annual Review of Earth and Planetary Sciences,29,535–562.
Aynekulu, E., & Shepherd, K. (2015). Trees for food security project-biophysicalbaselinedata,WorldAgroforestryCentre(ICRAF).Harvard Dataverse,V2,https://doi.org/10.7910/DVN/29652
Bates,D.,Maechler,M.,Bolker,B.,&Walker,S.(2015).Fittinglinearmixed-effectsmodelsusinglme4.Journal of Statistical Software,67,1–48.
Batterman,S.A.,Hedin,L.O.,vanBreugel,M.,Ransijn,J.,Craven,D.J.,&Hall,J.S. (2013).Keyroleofsymbioticdinitrogenfixation in tropicalforestsecondarysuccession.Nature,502,224–227.
Benayas, J. M. R., Newton, A. C., Diaz, A., & Bullock, J. M. (2009).Enhancementofbiodiversityandecosystemservicesbyecologicalres-toration:Ameta-analysis.Science,324,1121–1124.
Berendse,F.,vanRuijven,J.,Jongejans,E.,&Keesstra,S. (2015).Lossofplantspeciesdiversityreducessoilerosionresistance.Ecosystems,18,881–888.
Botta-Dukát,Z.(2005).Rao’squadraticentropyasameasureoffunctionaldiversity based on multiple traits. Journal of Vegetation Science, 16,533–540.
Bradford,M.A.,Jones,T.H.,Bardgett,R.D.,Black,H.I.,Boag,B.,Bonkowski,M.,…Lawton,J.H.(2002).Impactsofsoilfaunalcommunitycomposi-tiononmodelgrasslandecosystems.Science,298,615–618.
Braun-Blanquet,J. (1932).Plant sociology. InD.M.Raup&D.Jablonski(Eds.),The study of plant communities.NewYork,NY:McGraw-Hill
Burnham,K.P.,&Anderson,D.R.(2002).Model selection and multimodel infer-ence: A practical information-theoretic approach.NewYork,NY:Springer.
Cadotte,M.W.,Carscadden,K.,&Mirotchnick,N.(2011).Beyondspecies:Functionaldiversityandthemaintenanceofecologicalprocessesandservices.Journal of Applied Ecology,48,1079–1087.
Cardinale,B.J.,Duffy,J.E.,Gonzalez,A.,Hooper,D.U.,Perrings,C.,Venail,P.,…Naeem,S. (2012).Biodiversity lossand its impactonhumanity.Nature,486,59–67.
Chapin,F.S.(2003).Effectsofplanttraitsonecosystemandregionalpro-cesses:A conceptual framework for predicting the consequences ofglobalchange.Annals of Botany,91,455–463.
Chave,J.,Coomes,D.,Jansen,S., Lewis,S. L., Swenson,N.G.,&Zanne,A.E.(2009).Towardsaworldwidewoodeconomicsspectrum.Ecology Letters,12,351–366.
Chave,J.,Rejou-Mechain,M.,Burquez,A.,Chidumayo,E.,Colgan,M.S.,Delitti,W.B.,…Vieilledent,G.(2014).Improvedallometricmodelstoestimatetheabovegroundbiomassoftropicaltrees.Global Change in Biology,20,3177–3190.
Chazdon,R.L.(2008).Beyonddeforestation:restoringforestsandecosys-temservicesondegradedlands.Science,320,1458–1460.
Cleverly,J.R.,Smith,S.D.,Sala,A.,&Devitt,D.A.(1997).Invasivecapac-ityofTamarixramosissimainaMojaveDesertfloodplain:theroleofdrought.Oecologia,111,12–18.
deDeyn,G.B.,Cornelissen,J.H.C.,&Bardgett,R.D.(2008).Plantfunc-tional traits and soil carbon sequestration in contrasting biomes.Ecology Letters,11,516–531.
Díaz,S.,Lavorel,S.,deBello,F.,Quetier,F.,Grigulis,K.,&Robson,T.M.(2007). Incorporating plant functional diversity effects in ecosystemserviceassessments.Proceedings of the National Academy of Sciences of the United States of America,104,20684–20689.
Diemont,S.A.W.,Martin,J.F.,Levy-Tacher,S. I.,Nigh,R.B.,Lopez,P.R., & Golicher, J. D. (2006). Lacandon Maya forest management:Restoration of soil fertility using native tree species. Ecological Engineering,28,205–212.
Dregne,H.E.(2002).Landdegradationinthedrylands.Arid Land Research and Management,16,99–132.
DuránZuazo,V.H.,&RodríguezPleguezuelo,C.R.(2008).Soil-erosionandrunoffpreventionbyplantcovers.Areview.Agronomy for Sustainable Development,28,65–86.
Finerty,G.E.,deBello,F.,Bílá,K.,Berg,M.P.,Dias,A.T.C.,Pezzatti,G.B.,&Moretti,M.(2016).Exoticornot,leaftraitdissimilaritymodu-latestheeffectofdominantspeciesonmixedlitterdecomposition.Journal of Ecology,104,1400–1409.
| 67Journal of Applied EcologyLOHBECK Et aL.
Fornara,D.A.,&Tilman,D.(2008).Plantfunctionalcompositioninfluencesratesofsoilcarbonandnitrogenaccumulation.Journal of Ecology,96,314–322.
Garnier,E.,Cortez,J.,Billes,G.,Navas,M.-L.,Roumet,C.,Debussche,M.,…Toussaint,J.-P.(2004).Plantfunctionalmarkerscaptureecosystempropertiesduringsecondarysuccession.Ecology,85,2630–2637.
Grime,J.P. (1998).Benefitsofplantdiversitytoecosystems: Immediate,filterandfoundereffects.Journal of Ecology,86,902–910.
Grover,H.D.,&Musick,H.B.(1990).ShrublandencroachmentinSouthernNew Mexico, USA: An analysis of desertification processes in theAmericanSouthwest.Climatic Change,17,305–330.
Gyssels,G.,Poesen,J.,Bochet,E.,&Li,Y. (2005). Impactofplantrootsontheresistanceofsoilstoerosionbywater:Areview.Progress in Physical Geography,29,189–217.
Haggar,J.P.,&Ewel,J.J.(1997).Primaryproductivityandresourceparti-tioninginmodeltropicalecosystems.Ecology,78,1211–1221.
Handa,I.T.,Aerts,R.,Berendse,F.,Berg,M.P.,Bruder,A.,Butenschoen,O.,…Hattenschwiler,S.(2014).Consequencesofbiodiversitylossforlitterdecompositionacrossbiomes.Nature,509,218–221.
Hombegowda,H.C.,vanStraaten,O.,Köhler,M.,&Hölscher,D.(2016).Ontherebound:Soilorganiccarbonstockscanbouncebacktonearforestlevelswhenagroforests replaceagriculture in southern India.Soil,2,13–23.
Hooper,D.U.,Chapin,F.S.,Ewel,J.J.,Hector,A.,Inchausti,P.,Lavorel,S.,…Wardle,A.(2005).Effectsofbiodiversityonecosystemfunc-tioning:Aconsensusofcurrentknowledge.Ecological Monographs,75,3–35.
Iiyama,M.,Derero,A.,Kelemu,K.,Muthuri,C.,Kinuthia,R.,Ayenkulu,E.,… Sinclair, F. L. (2016). Understanding patterns of tree adoption onfarms in semi-arid and sub-humidEthiopia.Agroforestry Systems,91,271–293.
Ilstedt,U.,BarguesTobella,A.,Bazie,H.R.,Bayala,J.,Verbeeten,E.,Nyberg,G,…Malmer,A.(2016).Intermediatetreecovercanmaximizeground-waterrechargeintheseasonallydrytropics.Scientific Reports,6,21930.
Jackson,R.B.,Banner,J.L.,Jobbágy,E.G.,Pockman,W.T.,&Wall,D.H.(2002).Ecosystemcarbonlosswithwoodyplantinvasionofgrasslands.Nature,418,620–623.
vanKleunen,M.,Weber,E.,&Fischer,M.(2010).Ameta-analysisoftraitdifferencesbetween invasiveandnon-invasiveplantspecies.Ecology Letters,13,235–245.
Kourtev,P.S.,Ehrenfeld,J.G.,&Häggblom,M. (2002).Exoticplantspe-ciesalterthemicrobialcommunitystructureandfunction inthesoil.Ecology,83,3152–3166.
Lal,R.,Griffin,M.,Apt,J.,Lave,L.,&Morgan,M.G.(2004).Managingsoilcarbon. Science,304,393.
Laliberté, E., & Legendre, P. (2010). A distance-based framework formeasuring functional diversity from multiple traits. Ecology, 91,299–305.
Laliberté,E.,&Shipley,B.(2012).R-packageFD:Measuringfunctionaldi-versityfrommultipletraits,andothertoolsforfunctionalecology.
Lamb,D.,Erskine,P.D.,&Parrotta,J.A.(2005).Restorationofdegradedtropicalforestlandscapes.Science,310,1628–1632.
Laughlin,D.C. (2014).Applyingtrait-basedmodelstoachievefunctionaltargets for theory-driven ecological restoration. Ecology Letters, 17,771–784.
Lavorel, S.,Grigulis,K., Lamarque,P.,Colace,M.,Garden,D.,Girel, J.,…Douzet,R.(2010).Usingplantfunctionaltraitstounderstandtheland-scapedistributionofmultiple ecosystem services. Journal of Ecology,99,135–147.
Lebrija-Trejos,E.,Pérez-García,E.A.,Meave,J.A.,Poorter,L.,&Bongers,F.(2011).Environmentalchangesduringsecondarysuccessioninatropi-caldryforestinMexico.Journal of Tropical Ecology,27,477–489.
Lefcheck,J.S.(2015).piecewiseSEM:Piecewisestructuralequationmodel-inginRforecology,evolution,andsystematics.Methods in Ecology and Evolution,7,573–579.
Lin,D.,Anderson-Teixeira,K. J., Lai, J.,Mi,X., Ren,H.,&Ma,K. (2016).Traitsofdominant tree speciespredict local scalevariation in forestabovegroundandtopsoilcarbonstocks.Plant and Soil,409,435–446.
Lohbeck,M.,Poorter,L.,Martínez-Ramos,M.,&Bongers,F.(2015).Biomassisthemaindriverofchangesinecosystemprocessratesduringtropicalforestsuccession.Ecology,96,1242–1252.
Lorenz,K.,&Lal,R. (2005).Thedepthdistributionofsoilorganiccarbonin relation to landuseandmanagementand thepotentialofcarbonsequestrationinsubsoilhorizons.Advances in Agronomy,88,35–66.
Lorenz,K.,&Lal,R. (2014).Soilorganiccarbonsequestration inagrofor-estry systems. A review. Agronomy for Sustainable Development, 34,443–454.
Lüdecke,D.(2016a).sjPlot:Datavisualizationforstatisticsinsocialscience.Rpackageversion2.0.2.Retrieved fromhttps://CRAN.R-project.org/package=sjPlot
Lüdecke, D. (2016b). sjstats: Statistical functions for regression models.Rpackageversion0.3.0.Retrieved fromhttps://CRAN.R-project.org/package=sjstats
Madari,B.E.,Reeves,J.B.,Machado,P.L.O.A.,Guimarães,C.M.,Torres,E.,&McCarty,G.W.(2006).Mid-andnear-infraredspectroscopicassess-ment of soil compositional parameters and structural indices in twoFerralsols.Geoderma,136,245–259.
McCormack,M.L.,Dickie,I.A.,Eissenstat,D.M.,Fahey,T.J.,Fernandez,C.W.,Guo,D.,…Zadworny,M.(2015).Redefiningfinerootsimprovesunderstandingofbelow-groundcontributionstoterrestrialbiosphereprocesses.New Phytologist,207,505–518.
Minnemeyer, S., Laestadius, L., Sizer, N., Saint Laurent, C., & Potapov, P.(2011). Bonn challenge on forests, climate change and biodiversity. Washington,DC:SouthDakotaStateUniversity.
Nakagawa,S.,Schielzeth,H.,&O’Hara,R.B.(2013).AgeneralandsimplemethodforobtainingR2fromgeneralizedlinearmixed-effectsmodels.Methods in Ecology and Evolution,4,133–142.
Paquette, A., & Messier, C. (2011). The effect of biodiversity on treeproductivity: From temperate to boreal forests. Global Ecology and Biogeography,20,170–180.
Pereira,H.M.,Leadley,P.W.,Proenca,V.,Alkemade,R.,Scharlemann,J.P.,Fernandez-Manjarres,J.F.,…Walpole,M.(2010).Scenariosforglobalbiodiversityinthe21stcentury.Science,330,1496–1501.
Petchey,O.L.,&Gaston,K.J.(2006).Functionaldiversity:Backtobasicsandlookingforward.Ecology Letters,9,741–758.
Pimentel,D.,&Burgess,M.(2013).Soilerosionthreatensfoodproduction.Agriculture,3,443–463.
Pinheiro,J.,Bates,D.,DebRoy,S.,&Sarkar,D.&RCoreTeam(2016).nlme:Linear andNonlinearMixed EffectsModels. R packageversion 3.1-128.Retrievedfromhttp://CRAN.R-project.org/package=nlme
Prieto, I., Stokes, A., & Roumet, C. (2016). Root functional parameterspredict fine root decomposability at the community level. Journal of Ecology,104,725–733.
RCoreTeam(2014).R: A Language and Environment for Statistical Computing. Vienna,Austria:RFoundationforStatisticalComputing.
Reeves,J.B.III,Follett,R.F.,McCarty,G.W.,&Kimble,J.M.(2006).Cannearormid-infrareddiffusereflectancespectroscopybeusedtode-termine soil carbon pools? Communications in Soil Science and Plant Analysis,37,2307–2325.
Reynolds, T., Wassie, A., Wubalem, A., Liang, J., & Collins, C. (2016).EffectsofexoticEucalyptusspp.plantationsonsoilpropertiesinandaroundsacrednaturalsitesinthenorthernEthiopianHighlands.AIMS Agriculture and Food,1,175–193.
Sandel,B.,Corbin,J.D.,&Krupa,M.(2011).Usingplantfunctionaltraitsto guide restoration: A case study in California coastal grassland.Ecosphere,2,art23.
Scherer-Lorenzen,M. (2008). Functional diversity affects decompositionprocessesinexperimentalgrasslands.Functional Ecology,22,547–555.
Schroth,G.(1995).Treerootcharacteristicsascriteriaforspeciesselectionandsystemsdesigninagroforestry.InF.L.Sinclair (Ed.),Agroforestry:
68 | Journal of Applied Ecology LOHBECK Et aL.
science, policy and practice: Selected papers from the agroforestry sessions of the IUFRO 20th World Congress,Tampere,Finland,6–12August1995(pp.125–143).Dordrecht,TheNetherlands:Springer.
Stokes,A.,Atger,C.,Bengough,A.G.,Fourcaud,T.,&Sidle,R.C. (2009).Desirableplantroottraitsforprotectingnaturalandengineeredslopesagainstlandslides.Plant and Soil,324,1–30.
Tardif,A.,&Shipley,B. (2013).Usingthebiomass-ratioand idiosyncratichypotheses to predict mixed-species litter decomposition.Annals of Botany,111,135–141.
Thijs,K.W.,Aerts,R.,VandeMoortele,P.,Musila,W.,Gulinck,H.,&Muys,B. (2014). Contrasting cloud forest restoration potential betweenplantations of different exotic tree species. Restoration Ecology, 22,472–479.
Tremblay,A.,&Ransijn,J.(2015).Modelselectionandpost-hocanalysisfor(G)LMERmodels.Rpackageversion2.10.
UBS(2012).StatisticalAbstract.MbaleDistrictLocalGovernment,UgandaBureauofStatistics,Mbale,Uganda.
Vågen,T.-G.,&Winowiecki,L.(2014).NorthernRangelandsTrust.Baselineassessmentofrangelandhealth,KalamaandNamunyakconservancies.ICRAF,CIAT,Nairobi,Kenya.
Vågen,T.-G.,Winowiecki,L.A.,Abegaz,A.,&Hadgu,K.M.(2013).Landsat-basedapproachesformappingoflanddegradationprevalenceandsoilfunctionalpropertiesinEthiopia.Remote Sensing of Environment,134,266–275.
Vågen, T. G., Winowiecki, L. A., Robiglio, V., Okia, C., Lohbeck, M., &Cornelius,J. (2017).Thepotential for forest landscaperestoration indegradedfarmlands:Fillingknowledgegapsontherestorationofde-gradedsmallfolderlandscapemosaics.Harvard Dataverse,V1,https://doi.org/10.7910/DVN/EQFM2N
Vågen,T.-G.,Winowiecki,L.A.,TameneDesta,L.,&Tondoh,J.E.(2013).The Land Degradation Surveillance Framework (LDSF) field guide v3. Nairobi,Kenya:WorldAgroforestryCentre.
Vågen,T.-G.,Winowiecki,L.A.,Tondoh,J.E.,Desta,L.T.,&Gumbricht,T.(2016).MappingofsoilpropertiesandlanddegradationriskinAfricausingMODISreflectance.Geoderma,263,216–225
Warren,J.,Topping,C.J.,&James,P.(2009).Aunifyingevolutionarytheoryfor thebiomass–diversity–fertility relationship.Theoretical Ecology,2,119–126.
Wilsey, B. J., & Potvin, C. (2000). Biodiversity and ecosystem func-tioning: Importance of species evenness in an old field.Ecology,81, 887–892.
Winowiecki,L.,&Sinclair,F.(2016).RestorationofdegradedlandforfoodsecurityandpovertyreductioninEastAfricaandtheSahel:Takingsuc-cessesinlandrestorationtoscale.Harvard Dataverse,V5,https://doi.org/10.7910/DVN/SBL27O
Winowiecki, L., Vågen, T.-G., Massawe, B., Jelinski, N. A., Lyamchai, C.,Sayula, G., & Msoka, E. (2015). Landscape-scale variability of soilhealth indicators:Effectsof cultivationon soilorganic carbon in theUsambaraMountainsofTanzania.Nutrient Cycling in Agroecosystems,105,263–274.
Wood,S.A.,Karp,D.S.,DeClerck,F.,Kremen,C.,Naeem,S.,&Palm,C.A.(2015).Functionaltraitsinagriculture:Agrobiodiversityandecosystemservices.Trends in Ecology & Evolution,30,531–539.
Zak,D.R.,Holmes,W.E.,White,D.C.,Peacock,A.D.,&Tilman,D.(2003).Plant diversity, soil microbial communities, and ecosystem function:Arethereanylinks?Ecology,84,2042–2050.
Zhou,G.Y.,Morris,J.D.,Yan,J.H.,Yu,Z.Y.,&Peng,S.L.(2002).Hydrologicalimpacts of reafforestation with eucalypt and indigenous species: Acase study in southern China. Forest Ecology and Management, 167,209–222.
SUPPORTING INFORMATION
Additional Supporting Information may be found online in thesupportinginformationtabforthisarticle.
How to cite this article:LohbeckM,WinowieckiL,AynekuluE,OkiaC,VågenT-G.Trait-basedapproachesforguidingtherestorationofdegradedagriculturallandscapesinEastAfrica.J Appl Ecol. 2018;55:59–68. https://doi.org/10.1111/1365-2664.13017