In the Vakinankaratra region of the Malagasy Highlands, rainfed rice cultivation has developed rapidly, which has led to rapid degradation of soil fertility and reduced productivity. This study involved a comparison of rainfed rice in rotation with maize/Brachiaria(MZ-BR) or maize/soybean (MZ-SB) intercrops, managed by direct seeding (DS) or plowed with residue burial (PB) with conventional mineral fertilization (FM) or without fertilizer (F0). The umbric Acrisol (hyperdystric) presented an anion exchange capacity (AEC) and variable charge according to the pH. Mineralizable nitrogen (N-minz.) values decreased sharply from the surface to the deep horizons and became almost nil in the 90-120 cm horizon, while the N(minz.)contents of the DS treatments were systematically 5-7% higher than the PB treatment values for a given depth. The AEC of the 0-10 cm horizon was practically zero but increased significantly with depth and reached 2.58 cmol((c))kg(-1). AEC values were inversely correlated with the organic carbon (C-org.) content: AEC = 1.81-0.375 C(org.)The nitrate adsorption isotherms were linear and the nitrate partition coefficients (coef_nit) were correlated with the AEC values: coef_nit = 0.3363 AEC+0.1855 (R-2 = 0.954). The soil nitrate stock determinants were highlighted by the matrix of correlations of the principal component analysis: excheangeable acidity and aluminum, cation exchange capacity, Delta pH, C(org.)and clay and total silt contents were particularly important for the AEC and coef_nit. The different cropping systems showed no significant impact after 12 years on soil characteristics that could improve nitrogen bioavailability. We thus recommend using organic fertilization to raise the soil pH, C(org.)content and nitrogen bioavailability.

Dans le Vakinankaratra (1500m d'altitude) en climat tropical d'altitude (1300mm, 20°C), une etude visant a optimiser la nutrition azotee du riz pluvial d'une succession culturale biennale, Mais+soja/riz pluvial (R4), a partir de l'offre en azote du sol ferralitique sur alluvions volcano-lacustres d'Andranomanelatra, a ete conduite. Elle s'est attelee a quantifier l'offre reelle en azote mineral disponible dans le profil cultural et a estimer la demande de la plante, en cet element, au cours de son cycle cultural pour arriver a etablir le bilan entrees-sorties d'azote en des situations contrastees, se differenciant par l'âge de deux dispositifs (4 et 12 ans), d'une part et par les modes de gestion de sol (semis direct sous couverture vegetale, SD et labour, Lab.), d'autre part. Les entrees sont representees par l'azote fourni par les residus et l'azote mineralise dans le compartiment organique du sol. Les sorties quant a elles, sont constituees par la consommation en azote de la plante et les pertes de tout ordre, sortant du systeme. Les diverses mesures effectuees ont mis en evidence que les parcelles en SCV plus âge sont mieux pourvues en azote mineral que ceux recents tandis que celles en labour repete pendant plusieurs annees en sont plus appauvries que celles labourees plus recemment. Par ailleurs, independamment des modes de gestion du sol et de leur âge, en toute periode, les horizons de surface du profil cultural du riz pluvial de la succession culturale consideree presentent une offre moindre en azote mineral que les horizons de profondeur. Cette concentration en profondeur de l'azote est imputable aux caracteristiques specifiques de ce type de sol et en relation avec la variation inversement proportionnelle de la densite racinaire avec la profondeur. Si bien qu'en labour comme en SCV, le bilan entrees-sorties d'azote du dispositif de 4 ans n'est pas profitable a la plante cultivee. En effet, d'importantes pertes d'azote mineral par lessivage sont notees aussi bien en systeme avec labour qu'en SCV. (Resume d'auteur)

Share-ploughed tillage with residue removed (CT-R) is the traditional tillage practice in the Highlands of Madagascar. No-tillage with residue mulching (NT+R) is nowadays often used as an alternative cultivation practice. Soils (0–5cm layer) were sampled in Spring 2003 from both management systems after 11 years of soybean–maize annual rotation on a clayey Ferralsol. Soil aggregate stability can influence soil organic carbon (SOC) storage by its protection from microbial decomposition. The soil organic carbon (SOC) content was significantly impacted by systems and crop residues derived-carbon represented 64% of the annual benefit in SOC of NT+R system. The carbon associated with soil water stable macro- (200–2000μm), meso- (20–200μm) and microaggregates (<20μm) from both systems, and their physical protection was studied by an incubation experiment of intact vs. crushed aggregates. Results showed macroaggregate content was significantly higher in NT+R than in CT-R system and mesoaggregate content was significantly higher in CT-R than in NT+R. Macroaggregates associated-C were 1.8 time higher in NT+R than in CT-R (31.9 and 17.9gCg−1 soil, respectively) and made up the largest percentage (>80%) of the difference of SOC content between NT+R and CT-R systems. The amount of mineralized C over 28 days was higher in NT+R than in CT-R, and higher in meso- than in macroaggregates. However, crushing aggregates did not significantly affect the amount of mineralized C in macro- and mesoaggregates for both management systems. The macro- and mesoaggregates protected-C was lower than 54μgg−1 soil for both NT+R and CT-R systems. Hence, the physical protection of C in aggregate larger than 50μm was not the main process of C protection in the studied systems. Thus, C protection might occur in aggregates larger or smaller than 50μm via physico-chemical protection mechanisms by association of organic matter to clay and silt fractions, or by protection due to chemical composition.

In the Cerrado region of Brazil conventional soybean monoculture is since the 1980s being replaced by direct seeding mulch-based cropping (DMC) with two crops per year and absence of tillage practices. The objective of this study was to assess the long-term impact of DMC on soil organic matter accumulation and nitrogen (N) mineralization. Measurements of soil organic carbon (C) content, soil total N content and soil N mineralization, both under laboratory conditions using disturbed soil samples and under field conditions using intact soil cores were conducted on a chronosequence of 2-, 6-, 9- and 14-year-old DMC fields (DMC-2, DMC-6, DMC-9 and DMC-14, respectively). The average increase of organic C in the 0–30 cm topsoil layer under DMC was 1.91 Mg C ha−1 year−1. Soil total N increased with 103 kg N ha−1 year−1 (0–30 cm). The potential N mineralization rate under laboratory conditions (28°C, 75% of soil moisture at field capacity) was 0.27, 0.28, 0.39 and 0.36 mg N kg soil−1 day−1 for, respectively, the DMC-2, DMC-6, DMC-9 and DMC-14 soils. The corresponding specific N mineralization rates were 0.16, 0.15, 0.22 and 0.17 mg N g N−1 day−1. There was no obvious explanation for the higher specific N mineralization rate of soils under DMC-9, given the similar soil conditions and land-use history before DMC was introduced. Results from the in situ N incubation experiments were in good agreement with those from the laboratory incubations. We estimated that soil N mineralization increases with about 2.0 kg N ha−1 year−1 under DMC. The increase was mainly attributed to the larger soil total N content. These results indicate that even in the medium term (10 years), continuous DMC cropping has limited implications for N fertilization recommendations, since the extra soil N supply represents less than 20% of the common N fertilization dose for maize in the region.

L'estimation des flux de N2O est un enjeu essentiel pour evaluer l'effet des pratiques culturales sur la production de gaz a effet de serre serre. La quantification in situ des flux de N2O, resultant de 2 principaux processus biologiques, nitrification et denitrification, est delicate en raison d'une variabilite spatiale et temporelle elevee et de concentrations naturelles classiquement basses. Ce travail methodologique vise a : (i) montrer que l'approche classique qui consiste a estimer les flux de N2O a partir de mesures quotidiennes in situ peut conduire a omettre des flux importants sur des periodes courtes (peu d'heures), (ii) caracteriser les emissions potentielles de N2O du sol par nitrification et denitrification par des mesures de laboratoire et (iii) proposer une approche par la modelisation des emissions de N2O qui evite le probleme des mesures discretes. Cette derniere approche a consiste en la combinaison de 2 modeles : un modele mecaniste de transfert de l'eau et un modele de simulation des emissions de N2O par nitrification et denitrication. Deux situations agricoles tropicales de champ ont ete etudiees : (i) semis direct sur couverture vegetale et travail superficiel du sol sans restitution de residus en contexte de production de riz pluvial dans les Cerrados (Bresil central) ; (ii), semis direct sur les residus de culture et labour sans restitution de residu au sol sur les Hautes Terres malgaches. (Resume d'auteur)

Nitrous oxide (N2O) and methane (CH4) are important greenhouse gases (GHG) produced respectively by the naturally occurring microbial processes of incomplete denitrification or nitrification and methanogenesis. Tillage practices and climate affect the release of GHGs. No tillage (DMC) systems may increase CO2-C fixation in soil but also N2O and CH4 emissions. The aim of this article is to question whether the positive effect of a DMC system observed on carbon storage for the topsoil layer in Cerrado soils is offset or not by the N2O and/or CH4 emissions. Two 5-year-old systems, tillage (disc on the first 15 cm called offset: OFF) and a direct-sowing mulch-based crop system (DMC) with an additional cover crop were studied during a cropping cycle. N2O and CH4 fluxes are determined using a closed-chamber, N2O and CO2 concentrations are measured at 3 depths (10-, 20- and 30 cm). No significant difference between treatments was observed for both gases (for emissions and concentrations). Soil NO contents increase from surface to depth (30 cm) and range from 300 ppbv to 3 ppm for both treatments. Total annual estimated emissions of N2O range from 31 to 35 g N2O-N ha(-1) year(-1) for DMC and OFF respectively which is low and corresponds only to 0.03% of he total N-fertilizer applied. Monthly means N2O emissions were strongly correlated to monthly means of N2O content at 10 cm depth (R-2 = -0.66) and seem to increase exponentially with monthly mean Water Filled Pore Space WFPS (0-10 cm layer) (R-2 = 0.33). CH4 fluxes were very low as well: both treatment act as source of CH4 (245 and 403 g CH4-C ha(-1) year(-1) for DMC and OFF respectively. On a CO-C equivalent basis these results correspond to 4.1 and 4.7 kg CO2-C ha(-1) year(-1) for NO and to 1.9 and 3.1 kg CO2-C ha(-1) year(-1) for CH4 for DMC and OFF respectively. As a result, the carbon sequestration balance taking into account the CO2, CH4 and N2O on a CO2-C equivalent basis is in favour of DMC treatment considering that this treatment increases carbon storage originated from CO-C for the topsoil (0-10 cm) layer (350 kg C ha(-1) year(-1)) in comparison with OFF treatment. (C) 2007 Elsevier B.V. All rights reserved.

Soil organic matter and crop yields under contrasting agricultural practices : Long-term trial and carbon dynamics modelling in Sudano-Sahelian environment The unique long-term experiment of Saria (Burkina Faso) offers the possibility to analyse soil carbon dynamics over more than 40 years under different agricultural practices in the Sudano- Sahelian region. Through the study of various pract ices (organic and / or minerals fertilization, conventional or reduced tillage) there is a decreas e of carbon stocks compared to the non- deteriorated herbaceous fallow, after 10, 20 and 40 years of continuous cropping. The best practices allow to preserve 45 to 90% of the stocks observed under the initial fallow according to the duration of the experimentation. There is a relationship bet ween input quality and changes of soil carbon stocks. The importance of the 0-20 �m fraction and water soluble C is shown in the part icle-size fractionation. The definition of a critical value o f C in the soil, between 6 and 7 mg C g -1 is confirmed. The modelling of C stocks with the RothC model shows the existence of an unsuspected cause of loss or transfer from manure of about 67% which do not yield any impacts on the improvement of C stocks of the soil. Our hypothesis is that the soil macrofauna is an unsuspected agent of carbon transport. This finding suggests ad ditional processes to be included in carbon models. Increasing the storage of C in these soils is possible with the current farming practices but the economic profitability of the technical alterna tives is still unclear.

Soil organic matter (SOM) plays a central role in the functioning of ecosystems, and is beneficial from agronomic and from environmental point of view. Alternative cultural systems, like direct seeding mulch-based cropping (DMC) systems, enhance carbon (C) sequestration in agricultural soils and lead to an increase in soil macrofauna. This study aimed at evaluating in field mesocosms the effects of earthworms on SOM dynamics and aggregation, as influenced by residue quality and management.In the highlands of Madagascar, buckets were filled with 2 mm-sieved clayey Inceptisol. The effects of earthworm addition (Pontoscolex corethrurus), residue addition (rice, soybean, and no addition), and localization of the residues (mulched or buried) were studied. After 5 months, soil from mesocosms with earthworms had significantly lower C concentration and higher proportion of large water-stable macroaggregates (>2000 mu m) than those without earthworms, because of the production of large macroaggregates by earthworms. Earthworm effect on soil aggregation was greater with rice than with soybean residues. Casts (extracted from mesocosms with earthworms) were slightly enriched in C and showed significantly higher mineralization than the non-ingested soil (NIS), showing that at the time scale of our study, the carbon contained in the casts was not protected against mineralization. No difference in microbial biomass was found between casts and NIS.Complementary investigations are necessary to assess long-term effects of earthworm addition on SOM dynamics, the conditions of occurrence of physical protection, and the impact of earthworms on the structure of the microbial community. (C) 2007 Elsevier Ltd. All rights reserved.

Sous climats tropicaux humides a sub-humides, la gestion de l'azote pose souvent probleme pour une production agricole durable. Une etude realisee dans les Cerrados bresiliens a montre que les systemes SCV peuvent ameliorer la valorisation de cet element. Sur le long terme, les moindres pertes par erosion et la forte restitution de residus avec ces systemes contribueraient a une augmentation de 83 kg ha-1 an-1 des stocks de N total du sol. Ainsi, les SCV permettent apres 10 ans d'application d'augmenter de 25 kg ha-1 l'azote mineralise durant les 120 jours du cycle du mais. Sur le court terme, on constate que les residus se decomposent lentement en surface du fait de leur faible surface de contact avec la matrice sol, sans qu'il y ait d'immobilisation en debut de cycle par le pool microbien. Il y a parfois peu de differences liees a la nature de la plante intercalaire de couverture utilisee. Toutefois cette couverture peut jouer un grand role dans le recyclage de l'azote et diminuer ainsi les pertes totales par lixiviation de 25 a 50 %. En effet, ces pertes sont d'autant plus importantes qu'un important pool de N mineral est present dans le profil en debut de saison des pluies parce que non valorise par une plante de couverture en fin de cycle precedent. (Resume d'auteur)

The unique long-term experiment of Sana offers the possibility to analyze soil carbon dynamics over more than 40 years under different agricultural practices in the Sudano-Sahelian region. Complete historical datasets offer the possibility to test carbon models. Maintaining a high carbon level here appeared more difficult than predicted by current carbon models. This finding suggests additional processes to be included in carbon models. (Resume d'auteur)

The management and enhancement of soil organic carbon (SOC) is very important for agriculture (fertility) as well as for the environment (carbon (C) sequestration). Consequently, changes in soil management may alter SOC content. No-tillage (NT) practices are potential ways to increase SOC. We studied the SOC from agricultural soils in the Cerrados in Central Brazil. We compared two different tillage systems: conservation agriculture with no-tillage under cover crops (NT) and disc tillage (DT) for 5 years in a context of rainfed rice production. The soil is a dark red Oxisol with high clay content (about 40%). The objectives of the study were: (i) to evaluate the short-term (5 years) impact of tillage systems on SOC stocks in an Oxisol and (ii) to better understand the dynamics of SOC in different fractions of this soil. We first studied the initial situation in 1998, and compared it to the 2003 situation. NT with cover crop (Crotalaria) was found to increase the storage of C in the topsoil layer (0-10 cm) compared to DT. The difference observed for the 0-10 cm layer under NT in comparison with DT represented C enrichment under no-tillage amounting to 0.35 Mg C ha(-1) year(-1) and corresponding to less than 10% of cover crops residues returned to the soil. A particle-size fractionation of soil organic matter (SOM) showed that differences in total SOC between NT and DT mainly affected the 0-2 mu m fraction and, to a smaller extent the 2-20 mu m fraction.This specific enrichment of SOC in the silt and clay fraction was attributed to (i) the storage of a water soluble C in the field and (ii) the effect of soil biota and especially fauna activity. The mean residence time of carbon associated with the fine fractions being rather long, it might be assumed that the preferential storage in fine fractions resulted in a long-term carbon storage. This study suggests a positive short-term effect of a no-tillage system on C sequestration in an Oxisol. (c) 2006 Elsevier B.V. All rights reserved.

Stocker du carbone dans le sol permet d'ameliorer ses proprietes physico-chimiques et de reduire les emissions de dioxyde de carbone vers l'atmosphere. L'effet des systemes en semis direct avec couverture vegetale (SCV) sur le stockage de C dans le sol est etudie sur un dispositif agronomique de longue duree (11 ans) a Antsirabe, Madagascar (16°C, 1 300 mm). Quatre systemes sont etudies: un systeme en labour conventionnel avec exportation des residus de recolte [CT m/s, rotation mais (Zea mays L.)-soja (Glycine max. L.)], et trois systemes en SCV sans travail du sol, et avec restitution des residus de recolte [NT m/s, rotation mais-soja; NT m/m-d, rotation mais-mais avec une couverture vegetale de Desmodium uncinatum; et NT h/s-k, rotation haricot (Phaseolus vulgaris)-soja avec une couverture vegetale de Pennissetum clandestinum]. Le sol est tres argileux, a faible capacite d'echange cationique mais possedant des proprietes andiques pouvant influencer les potentialites de stockage du C du sol. A 0-5 cm, les teneurs en C sont plus elevees sous SCV (NT m/s, NT m/m-d et NT his) que sous labour (CT m/s), et a 5-10 cm, elles sont plus elevees sous NT m/m-d et NT m/s que sous NT h/s-k et CT m/s. Le stockage annuel de C, a masse de sol equivalente, est de 0,69 et 1,01 mg C.ha-1.an-1, sous NT m/s et NT m/m-d pour l'horizon equivalent a 0-20 cm, alors qu'il n'y a pas d'effet SCV observe pour l'horizon equivalent a 0-40 cm. Ceci peut etre du a la fois a l'absence reelle de stockage comme a une variabilite initiale des teneurs en C dans les horizons de profondeurs, car le labour n'est effectue que jusqu'a 20 cm de profondeur. Les differences de stockage de C entre NT et CT dans la couche 0-20 cm sont essentiellement attribuees aux quantites beaucoup plus importantes de residus organiques restituees par les systemes NT par rapport au systeme laboure CT, mais on peut aussi envisager qu'une partie de cette difference soit le fait d'une perte de C par erosion sous labour. Les teneurs en macroagregats stables (MA, 200-2 000 [mu]m) sont plus elevees sous NT m/s, NT h/s-k et NT m/m-d que sous CT m/s a 0-5 cm et a 5-10 cm. Cette teneur en MA est correlee positivement (R = 0,408, p < 0,05, n = 24) avec la teneur en C du sol, ce qui pourrait induire (i) une amelioration de l'agregation en fonction de l'augmentation de la teneur en C du sol et (ii) une protection du C se trouvant a l'interieur de ces agregats contre la mineralisation microbienne. Toutefois, la respirometrie ne montre pas une protection physique de C dans les sites de protection superieurs a 200 [mu]m pour NT m/s et CT m/s. Dans cette etude, le C stocke dans le sol pourrait alors etre protege contre la mineralisation par d'autres processus comme l'adsorption sur les colloides du sol ou la recalcitrance biochimique de la matiere organique du sol. (Resume d'auteur)