Carbon, nitrogen cycling and land cover changes during regrowth in African dry tropical forests: integrating perspectives from field and satellite data across a chronosequence in the Miombo Woodlands of western Tanzania

Abstract:

Seasonally dry tropical forests in Africa (SDTFs), such as forests in Miombo Woodlands, are experiencing high rates of deforestation, degradation and regrowth. Increasing proportions of forest are disturbed or composed of young regrowth stands (<40 yr), yet the degree and dynamics of how forest structure, biogeochemical and hydrological cycling recover with regrowth are poorly understood. Here, we examine how forest structure, carbon (C) and nitrogen (N) cycling change with regrowth following cultivation in forests of western Tanzania’s Miombo. This work addresses 3 questions: (1) What are the timescales of aboveground tree C stock recovery and patterns of soil mineral N availability with regrowth; (2) How does N demand for tree leaf production compare to indicators of available mineral N in surface soils from young to mature forest sites; (3) How does canopy structure vary with regrowth and disturbance and scale to Landsat-style satellite data? We established a chronosequence of 18 sites with ages 3 to >40 years since abandonment. At each, we inventoried trees to quantify aboveground tree C stocks, sampled soils to 100 cm to measure C, total and mineral N (NH4+, NO3-), and surveyed canopy cover with point-line transects, spherical densiometer and photometric leaf area measures. We also conducted soil incubations to determine nitrogen mineralization potentials. Tree C stocks ranged from 0.4 ± 0.1 Mg C ha^-1 for 3-4 year sites to 27.2 ± 5.2 Mg C ha^-1 for 30-40 year sites, and were 44.5 ± 7.4 Mg C ha^-1 for mature forest sites. Rates of aboveground tree C stock changes (0.78 – 0.89 Mg C ha^-1 yr^-1) were comparable to the few published for Miombo forests. However, tree C stocks at 10 - 24 year sites (5.2 ± 1.1 Mg C ha ^-1) were much lower than those reported in comparable studies. Only sites > 30-40 years had C stocks approaching mature forests. Further analyses will compare N dynamics from leaves and soil across the chronosequence, and relate them to the trends in tree C stocks. We use ground and canopy cover data to test remote sensing characterizations of land cover across disturbed and regrowth sites. Such scaling relationships will allow us to improve remote sensing characterization of land cover in African SDTFs and develop landscape-scale estimates of how forest cover changes affect C, N and water cycling regionally.