Litter from a tropical savanna woodland (miombo): Chemical composition and C and N mineralization

The decomposability of 15 litter types from four miombo woodland species, two tree species, Brachystegia spiciformis and Julbernardia globiflora, a grass species, Dactylotaenium aegyptium, and a sedge, Bulbostylis megastachys, was determined on the assumption that different plant components could be used as predictors for C and N mineralization. The chemical composition of the litter, in terms of lignin, polyphenol, cellulose and total C and N contents was determined. The C and N mineralization patterns of the litter were also investigated during 75 days of aerobic laboratory incubation with soil and correlated with the initial composition. The litter materials contained between 40–51% organic C and total N contents of 13 of the 15 litter types were low, ranging from 0.47% for D. aegyptium leaf to 1.29% for lichens + bark fragments found associated with B. spiciformis stem. The lignin contents ranged from 0.03% in lichens + bark fragments to 16.5% in twig litter from B. spiciformis. Polyphenol contents ranged from 0 to 6%, with the highest values being found in twigs and pods. Cumulative C mineralization data were fitted to first-order kinetics estimating the decomposition parameters. Seed litter from B. spiciformis evolved more CO2, 68.8% of initial C, than all the other litter materials tested. It was found that identical plant parts from the two tree species showed a similar rate of C mineralization, which was as follows: seed > lichens + bark fragments > tree leaflet litter > rachis litter > twig litter. Cellulose, which has received less attention in the literature, was found to influence both C and N mineralization patterns among other natural biopolymers (P < 0.001) but the content of ash-free available C, which was calculated as the difference between total C and lignin C + polyphenol C + cellulose C on an ash-free basis, had the highest R2 value (0.910) for C mineralization and total N had the highest R2 value (0.881) for N mineralization. Net N mineralization during 75 days was only apparent in 6 of the 15 litter samples, with results ranging from 0.2% of initial N in lichens + bark fragments from J. globiflora to 46% of initial N in B. spiciformis seed. The remaining samples immobilized N during the 75 days. Net N mineralization-immobilization was better correlated with the lignin C-to-N ratio (R2 = 0.695; P < 0.001) and the (lignin + polyphenol) C-to-N ratio (R2 = 0.688; P < 0.001) than the lignin-to-N ratio (R2 = 0.656; P < 0.001) or lignin + polyphenol-to-N ratio (R2 = 0.613; P < 0.001). Total N concentrations, C-to-N ratios and cellulose-to-N ratios appeared to be good predictors of N mineralization rates of the N-poor woodland (P < 0.001). The major part of miombo litter may not be a source for mineral N during the first 2–3 months after addition but the fractionation of litter into different C fractions gave a good indicator of potential C mineralization.