From Imperata cylindrica grasslands to productive agroforestry

Keywords: Ecosystem, Agroforestry, Imperata cylindrica , pioneer, mycorrhizae, inter-cropping, tree architecture, biomass, functional branching analysisConversion of an Imperata cylindrica ecosystem into an agroforestry ecosystem is a complex process. Integrated control of the I . cylindrica grass is needed in combination with planting deep-rooted pioneer tree species, a legume cover crop and annual food crops. I . cylindrica grasslands are generally poor in nutrients, but they are not necessarily poor from a soil biological perspective and may contain high densities of mycorrhizal spores. Experiments in the nursery and in the field were conducted to provide the knowledge needed to design and improve smallholder options for converting I. cylindrica grasslands into productive and sustainable agroforestry systems. Inoculation with mycorrhizal fungi (resting spores in 'Mycofer' belonging to Glomus manihotis , Glomus etunicatum , Gigaspora rosea and Acaulospora tuberculata ), did not promote tree seedling growth in the nursery, but it leads to a significant increase in survival rate of the four tree species tested, once these were transplanted to the field. However, this rate increased by 6.5% only. Nursery inoculation of the trees did not have any positive effect on the subsequent growth of those trees that survived the critical early stage. There is no compelling need for the use of the inoculum of the type used in the nursery. The performance of three out of four selected tree species growing in the alang-alang grassland, and their capability to outshade the grass, mahogany ( Swietenia macrophylla King, Meliaceae), sungkai ( Peronema canescens Jack., Verbenaceae) and candle nut ( Aleurites moluccana (L.) Willd., Euphorbiaceae) made them most suitable to be planted in those useless lands. Breadfruit or sukun ( Artocarpus altilis Fosberg, Moraceae) proved itself unsuitable for the harsh conditions of large parts of the I . cylindrica grasslands. It only performed well in specific sites. The field experiments showed no statistically significant differences in tree growth linked to the method used for initial alang-alang suppression. Ploughing, herbicide use or the simple 'pressing' method all proved acceptable as part of the overall system used. Spraying herbicide and ploughing are the most efficient methods to reduce alang-alang biomass. Pressing is the cheapest method to prepare acceptable tree growth, but is less attractive for inter-cropping. As a cover crop, able to outshade alang-alang after one year, to prevent its recovery and to improve soil fertility at the same time, Pueraria javanica (Leguminosae-Fabaceae) is a good option. An overall tree-LAI 1 and an overall rate of tree-shading 40 % are conditions for acceptable corn yields by inter-cropping in a matrix of diverse trees species. Mahogany and sungkai have a crown architecture and a root distribution apt to stimulate the establishment of an agroforestry system, because they are particularly suited to live together with annual plants. Candle nut, however, is more appropriate to reduce I. cylindrica grass by shading with its wide, dense hemispherical crown. Functional Branching Analysis (FBA) software can yield estimates of above ground biomass of mahogany and sungkai with an average standard deviation of 1 ± 0.34. While conversion of alang-alang grasslands into more productive and sustainable land uses is technically feasible, a broader approach including household level livelihood options may be needed to better understand the structure of the incentives needed to achieve real-world impacts of this research.