Abstract:

Soil macronutrient availability (particularly nitrogen (N) and phosphorus (P)) is a crucial abiotic
control for the cycling of carbon (C) and N in terrestrial ecosystems. However, empirical
evidence on macronutrient regulation of soil greenhouse gas (GHG; carbon dioxide (CO2), methane
(CH4), nitrous oxide (N2O)) and N leaching fluxes from tropical forests and agricultural
systems in sub-Saharan Africa (SSA) is still lacking. Yet, currently, SSA accounts for nearly
one-third of all tropical forests. It is also expected to become a hotspot for increased N deposition,
large-scale deforestation, and agricultural intensification in the near future. Hence, highresolution
measurements (spatially and temporally) on C and N fluxes from SSA terrestrial
ecosystems are needed to constrain global C and N budgets properly. Thus, this PhD study
evaluated the regulation effect of soil macronutrients on soil GHG and N leaching fluxes in a
nutrient-limited tropical forest and a fertilized sugarcane plantation in north-western Uganda.
The PhD study is based upon three interconnected work packages (WP).
In WPI, it is investigated how soil GHG fluxes (CO2, CH4, and N2O) were affected by macronutrient
limitations in a Ugandan tropical forest. Hence, a large-scale nutrient manipulation
experiment (NME) was setup in Budongo Forest Reserve consisting of four times replicated
plots with N, P, N + P, and control treatments. In every replicate plot, soil CO2, CH4, and N2O
fluxes were measured monthly (between May 2019 and June 2020) using static vented chamber
bases. The study findings show that N addition (N and N + P) resulted in significantly
higher N2O fluxes in the transitory phase (0-28 days (d) after fertilization). N fertilization likely
increased soil N beyond the microbial immobilization and plant nutritional demands, leaving
the excess to nitrification or denitrification. Prolonged N fertilization, however, did not elicit a
significant response in background N2O fluxes (measured more than 28 d after fertilization). P
fertilization marginally and significantly increased transitory and background CH4 consumption,
probably because it enhanced methanotrophic activity. Adding N and P together (N + P)
resulted in larger CO2 effluxes in the transitory phase, suggesting a possible co-limitation of
both N and P on soil respiration. Heterotrophic (microbial) CO2 effluxes were significantly
higher than the autotrophic (root) CO2 effluxes across all treatment plots, with microbes contributing
about two-thirds of the total soil CO2 effluxes. However, neither heterotrophic nor
autotrophic respiration significantly differed between treatments.
In WP2, it is assessed how forest conversion to intensively fertilized sugarcane plantations affected
soil GHG fluxes (CO2, CH4, and N2O). Here, soil GHG fluxes from the control plots in
WPI were compared to those measured in every treatment plot of a completely randomized
design (CRD) experiment in a sugarcane plantation. The CRD experiment was established in a
iii
Summary
iv
5.6-hectare ratoon sugarcane field 6 km south of the forest NME. It consisted of fertilizer treatments
(low, standard, and high) that represented a gradient of N fertilization rates used by sugarcane
farmers in north-western Uganda. Similarly (like in the NME), all the CRD treatments
were replicated four times. Soil GHG fluxes were determined with static vented chambers intensively
in the six months that followed fertilization before switching to monthly measurements
for the remaining period of the sampling campaign. Additionally, for every land use, fine
root biomass was determined based on 20 x 20 x 10 cm soil monoliths while soil organic carbon
(SOC) stocks were determined based on oven-dry bulk densities and SOC concentrations in the
first 1-meter soil depth. Soil CO2 effluxes were higher under sugarcane compared to the forest
because of the higher autotrophic respiration from the sugarcane’s fine root biomass and the
microbial decomposition of the sugarcane’s larger SOC stocks. Conversely, soil CH4 uptake
under sugarcane was three times lower than under forest, owing to the likely alteration of methanotroph
abundance upon conversion. Likewise, soil N2O emissions were much smaller under
sugarcane than in the forest because excess N from fertilizer addition in the sugarcane was
either lost through leaching or taken up by the sugarcane crop. All the results combined demonstrate
that even with the higher soil CO2 effluxes under sugarcane compared to the forest, the
fact that there was higher SOC sequestration in sugarcane plantations of different ages relative
to the native forest, suggests that sugarcane systems in the study area acted as a C sink since
the uptake of CO2 far exceeds SOM mineralization. However, the SOC sequestration under
sugarcane does not offset the initial significant loss in the above and belowground biomass C
loss immediately after forest conversion. Moreover, the C sink under sugarcane can change if
CO2-equivalents related to N2O and CH4 fluxes a considered in the calculation of the sugarcane’s
C footprint.
In WP3, it is evaluated how increasing N fertilization rates affected N dynamics, productivity,
and profitability of sugarcane plantations established on Ferralsols. Here, soil N2O fluxes from
WP2 were used in combination with the measured N leaching fluxes and field fresh weight
(yield/biomass) from the respective treatment plots of the CRD experiment established in WP2.
N leaching fluxes were determined based on drainage fluxes estimated with the Leaching Estimation
and Chemistry Model and leachate N concentrations obtained from suction cup lysimeters
installed at the soil depth of 90 cm. However, estimation of N leaching fluxes was limited
by the lack of site-specific measurements of soil hydraulic properties and pedotransfer functions
(PTFs) trained and calibrated for Ferralsols in tropical Africa. This challenge was overcome by
testing a suite of American, Brazilian, and European PTFs for their suitability in determining
soil hydraulic properties for the study test site. Sugarcane field fresh weight was estimated by
Summary
v
randomly harvesting four (1 m x 1 m) quadrants in every replicate plot. In WP3, it is demonstrated
that three of the five tested PTFs reliably estimated drainage fluxes for the study test site
in Uganda based on the match between the measured and predicted soil matric water potentials.
Therefore, despite the tested PTFs being developed using American, Brazilian, and European
soil datasets, some of them were robust enough to be used outside their training and validation
geographical confines with a satisfactory degree of accuracy. N leaching fluxes marginally increased
when N rates were increased from low to standard but significantly when the N rates
exceeded the standard rate. The measured soil N?O emissions were unaffected by N fertilization.
Sugarcane yields did not respond to increasing N rates, despite a significant to marginal
increase in crop N uptake between low and standard N rates and at N rates higher than the
standard, respectively. All the findings from WP3 suggest that surpassing the standard N rate
for sugarcane in north-western Uganda would be less economically viable since it would only
marginally increase yields, while the substantial increase in N leaching will affect groundwater
quality. Additionally, despite demonstrating that sugarcane cultivation can still be profitable at
lower-than-standard N rates since part of the N requirement is met by mineralizing the high soil
organic matter levels in sugarcane fields, it remains unreconciled from this short-term study
whether reducing N rates below the standard N rate will not counterintuitively lower SOC
stocks in the long term. The high SOC stocks under sugarcane reflect the long-term C input
dynamics obtained with the standard N rates.