Trades-offs between biomass use and soil cover. The case of rice-based cropping system in the lake Alaotra region of Madagascar.

International audienceFarmers in the Lake Alaotra region of Madagascar are currently evaluating a range of conservation agriculture (CA) cropping systems. Most of the expected agroecological functions of CA (weed control, erosion control and water retention) are related to the degree of soil cover. Under farmers' conditions, the grain and biomass productivity of these systems is highly variable and the biomass is used for several purposes. In this study,we measured biomass production of cover crops and crops in farmers' fields. Further, we derived relationships to predict the soil cover that can be generated for a particular quantity of mulch. We used these relationships to explore the variability of soil cover that can be generated in farmers' fields, and to estimate howmuch of the biomass can be removed for use as livestock feed,while retaining sufficient soil cover. Three different kinds of cropping systems were investigated in 91 farmers' fields. The first two cropping sequences were on the hillsides: (i) maize + pulse (Vigna unguiculata or Dolichos lablab)inyear1, followed by upland rice in year 2; (ii) the second crop sequence included several years of Stylosanthes guianensis followed by upland rice; (iii) the third crop sequence was in lowland paddy fields: Vicia villosa or D. lablab, which was followed by rice within the same year and repeated every year. The biomass available prior to rice sowing varied from 3.6 t ha−1 with S. guianensis to 7.3 t ha−1 with V. villosa. The relationship between the mulch quantity (M) and soil cover (C) was measured using digital imaging and was well described by the following equation: C = 1 − exp(−Am × M),where Am is an area-to-mass ratio with R2 > 0.99 in all cases. The calculated average soil cover varied from 56 to 97% for maize + V. unguiculata and V. villosa, respectively. In order to maintain 90% soil cover at rice sowing, the average amount of biomass of V. villosa that could be removed was at least 3 t ha−1 for three quarters of the fields. This quantity was less for other annual or biennial cropping systems. On average the V. villosa aboveground biomass contained 236 kg N ha−1. The study showed that for the conditions of farmers ofMalagasy, the production and conservation of biomass is not always sufficient to fulfil all the above-cited agroecological functions of mulch. Inventory of the soil cover capacity for different types of mulch may help farmers to decide how much biomass they canremove from the field