Above-ground earthworm casts affect water runoff and soil erosion in Northern Vietnam

Above-ground earthworm casts affect water runoff and soil erosion in Northern Vietnam
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Above-ground earthworm casts affect water runoff andsoil erosion in Northern Vietnam Pascal Jouquet   a,b, ⁎ , Pascal Podwojewski  b,c , Nicolas Bottinelli  a,b , Jérôme Mathieu  a  ,Maigualida Ricoy  d , Didier Orange  b,c , Toan Duc Tran  b , Christian Valentin  c a   IRD, UMR 137 Biosol, 32 Avenue H. Varagnat, 93143 Bondy Cedex, France  b SFI - IRD - IMWI, Dong Ngac, Tu Liem, Hanoï, Vietnam c  IRD, UR176 SOLUTIONS, 32 Avenue H. Varagnat, 93143 Bondy Cedex, France d Universidad de Vigo, Facultad de Biologia, Departamento de Ecologia y Biologia Animal, Campus as Lagoas-Marcosende, 36310 Marcosende, Spain Received 5 July 2007; received in revised form 7 December 2007; accepted 20 December 2007 Abstract This manuscript focuses on the effects of above-ground earthworm casts on water runoff and soil erosion in steep-slope ecosystems in NorthernVietnam. We investigated the effects of   Amynthas khami , an anecic species producing above-ground casts of prominent size, on water infiltrationand soil detachment along a land-use intensification gradient: a cultivation of cassava (  Mahinot esculenta ; CAS), a plantation of Bracharia(  Bracharia ruzziziensis ; BRA), a fallow (FAL), a fallow after a forest of   Eucalyptus  sp. (EUC) and a plantation of trees (  Acacia mangium  and Venicia Montana ; FOR). Two scales of studies were considered: (i) at the structure scale (cm 2 ), a water runoff simulation was used todifferentiate the effects of casts, free biogenic aggregates that previously belong to casts, and free physicogenic aggregates; (ii) at the station levels,1-m 2  plots were used to determine runoff and soil detachment rates during the rainy season in 2005.  A. khami  was sensitive to land-use management. Earthworm density was low in all the fields (0 – 1 ind m − 2 ). The highest densities were found inEUC and FOR and no individual was found in CAS. As a consequence, soil surface in EUC and FOR was covered with casts and free biogenicaggregates(approximately22and8kgm − 2 ,respectively).InFALandBRA,castscoveredthesoilonlysparselywith b 3kgm − 2 .InCAS,soilsurfacewascharacterized byfreephysicogenicaggregates thatmightbeproducedbyhumanactivityorendogeic earthwormsthroughtillage(approximately1 kg m − 2 ). Water runoff simulation clearly showed an enhancement of water infiltration with earthworm casting activity. Water runoff was moredecreasedwithcasts(  R 2 =0.26)thanfreebiogenicaggregates(  R 2 =0.49).Conversely,physicogenicaggregateswerenotassociatedwithhigherwater infiltration. Analyses of runoff and soil detachment rates during the rainy season underlined that the more land-use type have aggregates on soilsurface and the less important is surface runoff (  R 2 =0.922). Conversely, no relation occurred between aggregates and soil detachment rate. Whileabove-ground casting activity decreased surface runoff, they were not involved in soil detachment, and therefore soil erosion.© 2007 Elsevier B.V. All rights reserved.  Keywords:  Erosion; Earthworms; Above-ground casts; Land-use change; Water runoff; Soil detachment  1. Introduction Soil erosion is a widespread land degradation problem at theglobal scaleintermofloss ofsoil fertility andwater quality(Lal,2004, 2005). Land-use change, with the loss of the protectivevegetationcover,isoftenconsideredasthemainhumanfactorof soil erosion. In South-East Asia, erosion is regarded as a major typeofenvironmentaldamage(MaglinaoandLeslie,2001).Dueto rapid human population growth, the cropping areas haveexpanded to more marginal lands such as mountains and thefallowperiodshavebeenshortenedorevenabandoned(Clement et al., 2006). In Northern Vietnam, much of the rain forest in themountain was lost during the 1970s, and the trend continues(Sharma, 1992; Castella et al., 2006). Forests were cut to expand cultivated cassava, arrowroot, taro, maize and  Eucalyptus cropping on the uplands. Due to decreased soil fertility in theuplands, farmers have gradually converted their plots under annual crop cultivation into common grazing land, or tree  Available online at Catena 74 (2008) 13 – ⁎  Corresponding author. IRD, UMR 137 Biosol, 32 Avenue H. Varagnat,93143 Bondy Cedex, France.  E-mail address:  (P. Jouquet).0341-8162/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.catena.2007.12.006   plantations (mainly acacias) (Tran Duc et al., 2004; Clement et al., 2006).Land-usechangeisoftenaccompaniedwithalossandshiftof soilmacrofaunadiversity.Usuallyconsideredasbeingoneofthemost important macrofaunal organisms in soil, earthworms arevery sensitive to land-use change (Paoletti, 1999; Curry et al.,2002). Any shifts in earthworm community might havesignificant consequences in term of soil erosion in steep-slopetropical ecosystems. The influence of earthworm on soil properties is species-specific. Through their burrowing andcasting activities, earthworms living both below and above-ground (anecic and epi-anecic species, sensu Bouché, 1977)significantlyaffectsoilsurfacepropertieswithinthetoplayersof soil (Lavelle and Spain, 2001). In temperate ecosystems withgentle slopes, casts produced on the soil surface increase soilroughnessandinturnaffectwaterrunoffvelocityandinfiltrationintosoil(BinetandLeBayon,1999;LeBayonandBinet,2001).Galleries connected with the soil surface might also constitute preferential flow paths for water infiltration (Bastardie et al.,2003, 2005; Chan, 2004). However, beneficial effects of theseearthwormspeciesonsoilerosionmightbeoffsetbythelowsoilstructural stability of their casts when they are freshly emitted(Shipitalo and Protz, 1988; Le Bayon and Binet, 1999). Indeed, Le Bayon et al. (2002) showed that earthworm casts might increase soil detachment and nutrient transfers during rainstormevents in Brittany (North-West France) (Le Bayon et al., 2002). Conversely, earthworm species that live below-ground do not affect soil surface properties and are assumed to play a lesssignificant role. Surprisingly, the influence of earthwormdiversity and activity on water runoff and soil erosion have been only poorly studied in sloping lands of the tropics, whichhave been identified as one of the most biogeochemical activecycling in the world (Koch et al., 1995).In the mountains of the Northern Vietnam,  Amynthas khami  builds water-stable casts that are deposited on the soil surface.These biogenic structures can reach 20 cm height and areformed after week of daily deposition of globular casts at the topedge of the structure. These casts can be broken, probably bylivestock trampling and human walking. Hence, free biogenicaggregates can be released on the soil surface and constitute asignificant quantity of free macro-aggregates on the soil surface.This study aims at determining the effect of this accumulation of casts and free biogenic aggregates on surface runoff and soildetachment within steep-slope ecosystems. Two experimentswere set up along a land-use intensification gradient: (i) a water runoff simulation was carried out at the cm 2 scale to determinethe effect of earthworm casting activity on water runoff andinfiltration, and (ii) infiltration plots (1-m 2 ) were set up todetermine if above-ground cast production is associated withsoil detachment, and therefore soil erosion. 2. Materials and methods 2.1. Study site The experimental catchment of the MSEC (Managing SoilErosion Consortium of International Management of Water Institute, IMWI) project of Dong Cao (46 ha) is located in north-east Vietnam, approximately 50 km south-west of Hanoi (20°57 ′  N, 105° 29 ′ E). This experimental catchment is followedsince 1999 for the measurement of water fluxes, runoff anderosion rates in relation with land-use change. This watershed issurrounded by hills with a general slope of 40% in average but sometimes reaching 100%. The annual rainfall ranges from1500 – 1800 mm, of which 80 – 85% is concentrated from Aprilto October. The air humidity is always high, between 75 and100%. The mean daily temperature varies from 15 °C to 25 °C(Tran Duc et al., 2004).The dominant soil type is an Acrisol (WRB, 1998) or Ultisol(SSS, 1999). Soils derived from the weathering of volcano-sedimentary schists of Mesozoic age. Soils are over 1.0 m deep but with marked variation in depth. They have more than 50%claycontent,mainlykaolinitewithalowCEC( b 10cmolkg − 1 ),and are very porous with a bulk density of 1000 kg m − 3 . Theyhave a homogenous brown colour 10YR4/4 to 7.5YR 4/6, and aweak vertical differentiation.Before the 1960s, the region was covered with a dense primary forest. Deforestation and conversion to agriculturalland has led to its disappearance. From the mid 70s until veryrecently, villagers have cultivated cassava, taro and maize on theuplands (Castella et al., 2006). Since 1998, the catchment wascovered mainly with cassava with some area of   Eucalyptus  plantation and the previous crop was Maize. From 2002 practices changed very quickly because of the soil erosion andsoil fertility decrease. Annual soil loss recorded through bedload measurements have decreased from 3.6 t ha − 1 year  − 1  before 2002 to 0.1 – 0.3 t ha − 1 year  − 1 in 2004 (Orange et al.,2007). Finally, five agro-systems were dominant in 2005: (1) ayoung fallow after 4 years of cassava from 2002 (FAL), (2) a plantation of   Acacia mangium  and  Venicia montana  planted in2001 after a cassava plantation (FOR), (3) a fodder plantationwith  Bracharia ruzziensis  planted in 2003 after cassava (BRA),(4) a very young regrowth of   Eucalyptus  sp. trees, following a  Eucalyptus  plantation cut in 2003 (EUC), and (5) a small area of cassava (  Manihot esculenta ) (CAS). These five land usesrepresent the diversity of managements in this region of Southeast Asia. No special treatment has been applied in theseagro-systems, even under cassava (no herbicide, low fertiliza-tion, just superficial tillage and weeding before cultivation inCAS and BRA). Vegetation cover on the ground was low under FOR (few small shrubs, lot of leaves). There was no litter inCAS and very few in FAL and BRA. FAL and BRA were verysimilar in term of vegetation and litter cover. Under EUC, litter and vegetation cover were more important and shrubs growwith a large density. 2.2. Sampling of earthworms, surface casts and soil aggregates Earthworms were hand-sorted from 1 m×1 m×50 cm deepmonoliths. Sampling was done during the rainy season (August 2005) when communities were assumed to be at peak of abundance. Soil samplings were randomly repeated 10 times ineach land-use type. Earthworms were rapidly hand-sorted after soil excavation and identified at the species level. In this study, 14  P. Jouquet et al. / Catena 74 (2008) 13  –  21  the overall density of earthworms and the specific abundance of   A. khami  were considered. This earthworm species is con-sidered as anecic sensu Bouché (1977) and create verticalgalleries until more than 60 cm deep. Its size is highly variableand individuals can probably reach more than 50 cm long at theadult stage.Above-groundsoilmacro-aggregates N 5mmwerecollectedinthe different land-use types in 25×25 cm plots ( n =9, Fig. 1).AccordingtothedefinitionsdescribedinBullocketal.(1985)andPulleman et al. (2005), we distinguished three groups:(i) epigeous casts (fresh casts: CAST), (ii) free biogenic aggre-gates (rounded shape macro-aggregates that were clearly iden-tified as belonging to old casts: ROUND) and (iii) physicogenicaggregates (angular to subangular blocky macro-aggregates:ANG). The morphological fractions were air-dried and weighedto determine their relative mass contribution. Fig. 1. Examplesofsoilmacro-aggregates:(a)CAST:castsdepositedby  Amynthaskhami  and (b) ROUND: free rounded shape aggregate clearly identified has belonging to old cast (i.e. biogenic aggregates); (c) ANG: (sub)angular blockyaggregate (i.e. physicogenic aggregate) (Photos Pascal Jouquet, IRD).Fig. 2. Water runoff simulation: 1.5 L Methylene blue water was added in aconstant flow (75 mL s − 1 ) on the top of a 50 ⁎ 30 cm Plexiglas plate. Water reached the soil with a uniform front of 30 cm width along a 40% slope.Measured parameters were soil moisture in the surrounding environment,distance and surface covered by runoff water (determined from dyed soil),surface occupied by casts (CAST), biogenic aggregates (ROUND) and physicogenic aggregates (ANG) and vegetation.Table 1Overall density of earthworms and specific density of   Amynthas khami (ind m − 2 ) in the different land use (CAS: cassava plantation; FAL: fallow;BRA: bracharia plantation; EUC: fallow and  Eucalyptus  regrowth; FOR: forest)(Mean±standard error,  n =10)Density (ind m − 2 )Total  Amynthas khami CAS 4.29 (±0.70) 0.00 (±0.00)FAL 13.50 (±2.03) 0.25 (±0.44)BRA 9.60 (± 1.38) 0.30 (± 0.42)EUC 6.99 (±1.13) 1.11 (± 0.59)FOR 4.03 (± 0.57) 0.63 (± 0.78)15  P. Jouquet et al. / Catena 74 (2008) 13  –  21  2.3. Water runoff simulation Water runoff simulation was carried out by adding 1.5 Lmethylene blue water ( ≈ 5 gL − 1 ) with a constant flow of 75 mLs − 1 on the top of a 50 ⁎ 30 cm Plexiglas plate (Fig. 2). The platelayonthegroundandwaterreached thesoilwithauniformfront of 30 cm width. Given the pronounced impact of slope gradient onrunoffanddetachment(Janeauetal.,2003),asameslopewasset,around40%,whichisthemeanslopevalueofthecatchment.Litter was carefully removed from the soil surface to avoid anyinfluence on water runoff. Methylene blue water dyed the soiland allowed us to immediately and precisely determine thedistance and surface covered by water runoff. Different parame-ters were measured to describe soil surface properties. Surfaces(%) covered by macro-aggregates (CAST, ROUND and ANG)and vegetation were visually estimated in a 25 ⁎ 25 cm frame positioned3cmdownslopefromthePlexiglasplate.Themaximumdistance covered by water runoff (cm) was measured while theoverall area covered by water runoff (cm 2 ) was visually estimated.The ratio distance:surface was used as an index to describe theshape of runoff (i.e. linear surface  vs  sheet surface runoff). Soilmoisture was measured in the surrounding environment.This experiment was done in each land use in differentiatingareas without macro-aggregates (control plots: CAS, FAL andBRA,  n =5), to areas with CAST and ROUND aggregates(FALcast,  n =6; BRAcast,  n =10; FORcast,  n =12; EUCcast, n =6),areaswithonlyROUNDaggregates(FALround,BRAround,FORround, EUCround,  n =6) or only physicogenic aggregates(CASang,  n =4) on the soil surface. 2.4. Runoff plots Three 1-m 2  plots were set up in each land-use type: CAS,FAL, BRA, EUC and FOR. The slope ranged from 39 to 44%, Table 2Weight of air-dried macro-aggregates on soil surface (kg m − 2 ): casts (CAST),free rounded shapes biogenic aggregates (ROUND) and free angular andsubangularaggregates(ANG) inthe different landuse (CAS:cassavaplantation;FAL: fallow; BRA: bracharia plantation; EUC: fallow and  Eucalyptus  regrowth;FOR: forest) (Mean±standard error,  n =9)CAST ROUND ANGCAS 0.00 (±0.00) 0.00 (±0.00) 0.93 (±0.59)FAL 0.88 (±1.18) 2.15 (±1.74) 0.31 (± 0.28)BRA 0.11 (± 0.11) 0.82 (± 1.03) 0.20 (± 0.28)EUC 11.57 (± 2.80) 11.24 (± 4.42) 0.00 (± 0.00)FOR 4.42 (± 1.10) 3.27 (± 0.77) 0.00 (± 0.00)Fig.3. Principalcomponentsanalysis(PCA) on the waterrunoff simulationin the different landuse (CAS:cassava plantation; FAL: fallow; BRA: brachariaplantation;EUC: fallow and  Eucalyptus  regrowth; FOR: forest): distance (cm) and surface (cm 2 ) covered by coloured runoff dyed water, soil moisture (%), and soil cover  properties (% cover of herbs, % aggregates: casts (CAST), biogenic- (ROUND) and physicogenic aggregates (ANG) in a 25 ⁎ 25 cm square beneath the Plexiglass plate). Control plots without free macro-aggregates on soil: CAS, FAL, BRA; Plots with CAST and ROUND: FALcast, BRAcast, FORcast, EUCcast; Plots withROUND aggregates: FALround, BRAround, FORround, EUCround; Plots with ANG aggregates: CASang. (a) Correlation circle. (b) Ordination of the samples in the plane defined by axes 1 and 2 of the PCA.16  P. Jouquet et al. / Catena 74 (2008) 13  –  21  representing the average slope value of the catchment. Plotswere bounded by rigid metal frames inserted to a depth of 0.10 m. Soil detachment and water runoff were collected after each rainfall event, from May to October 2005, in a collector at the outlet of the plot, as described by Janeau et al. (2003).Runoff rate was determined by the ratio of the quantity of water that runoff in the collector tank to the amount of daily rainfall.Soil detachment rate was determined through the measurement of sediment weight after filtration from runoff water and heatingat 105 °C. This sediment weight is assumed to represent thequantity of soil losses during the rainfall event on 1-m 2  plots.Surfaces occupied by macro-aggregates (CAST+ROUND+ANG) on these plots were visually estimated in October 2005,which corresponds to the end of the rainy season. 2.5. Statistical analyses Prior to analyses data were inspected for homogeneity of variance using the Levene's test and log-transformed whenrequired. Differences between treatments were tested throughanalysis of variance (ANOVA). Principal component analysis(PCA) was done using a matrix of 58 samples and 7 variables(distance and surface of water runoff, percentage of herb,CAST, ROUND and ANG cover, and soil moisture). Allstatistical calculations were carried out using  R  (R Development Core Team, 2004). Differences were considered significant,only when  P   values were lower than 0.05. 3. Results 3.1. Abundance of A. khami and quantity of surface aggregates The abundance of earthworms and more specifically of   A. khami  was site-specific (Table 1). The density of earthwormwas low and earthworms were mainly endogeic species (non- pigmented earthworms). Earthworms were more abundant inFAL as compared to the other agro-ecosystems.  A. khami  wasnot find in CAS and densities were very low in BRA and FAL( b 1 ind m − 2 ), without any significant difference between them(  P  N 0.05). Although the density of   A. khami  in FOR was higher than in FAL and BRA and lower than that in EUC, the resultswere not significantly different (  P  N 0.05 in all the cases). Thedensity of   A. khami  was the highest in EUC (approximately1 ind m − 2 ) and significantly different from that in FAL andBRA (  P  =0.011 and 0.012, respectively).As a consequence, we did not find any biogenic aggregates,either CAST or ROUND, in CAS (Table 2). The quantity of CAST in FOR was greater than in FAL and BRA (  P  b 0.001 inall the cases) and lower than in EUC (  P  b 0.001). Conversely, nosignificant difference occurred betweenthe quantity of ROUNDin FAL, BRA and FOR (  P  N 0.05) and the highest value wasfound in EUC (  P  b 0.001 in all the cases). CAS was charac-terized by the highest quantity of ANG aggregates (  P  b 0.005 inall the cases) and no significant difference occurred betweenFAL, BRA, EUC and FOR (  P  N 0.05). Fig. 5. Distance:surface runoff ratio in the different land use (CAS: cassava plantation;FAL:fallow;BRA:brachariaplantation;EUC:fallowand  Eucalyptus regrowth; FOR: forest). Control: bare soil; ANG or ROUND: soil covered by physicogenic and biogenic macro-aggregates, respectively; CAST: soil withearthworm casts on the soil surface (bars indicate standard errors).Fig. 4. The relationship between casts (CAST), biogenic (ROUND) and physicogenic (ANG) aggregates (% soil cover) and dye stained area (cm 2 )( n =66). Linear regression lines describing the relationship between soilaggregates and dye stained area are fitted (  y = − 5.28  x +757.30,  R 2 =0.490,  P  b 0.001 for ROUND;  y = − 7.42  x +595.73,  R 2 =0.259,  P  b 0.001 for CAST;  R 2 =0.01,  P  =0.198 for ANG).17  P. Jouquet et al. / Catena 74 (2008) 13  –  21
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks