Bjorn Dirks

This thesis begins from the observation that the increase in the atmospheric CO2 concentration has not been increasing as fast as could be expected on basis of anthropogenic CO2 emissions. Close examination shows that the terrestrial biosphere exhibits a large annual CO2 sink of approximately 12.5 Gt CO2. This equates approximately 50% of anthropogenic CO2 emissions, which implies that without this biospheric sink the atmospheric CO2 concentration could have increased to well over 500 μmol mol-1 instead of its current 410 μmol mol-1. The study aims to explore to which extent the grassland biome could play a role in this biospheric net CO2 sink. It starts by analysing how a CO2 balance in grassland ecosystems emerges from its constituent processes. Its successive chapters develop how diurnal cycles of assimilatory and respiratory activity aggregate to an annual cycle of CO2 exchange between atmosphere and grassland ecosystem. It establishes why the annual net CO2 exchange or CO2 balance differs strongly among years.

Particular attention in this analysis is given to drained peat grassland ecosystems. Undrained peat grasslands and associated types of wetlands used to be very common and are characterised by large amounts of C immobilised in an anaerobic soil profile. Many of these peat grasslands have been drained to improve agricultural productivity and now have a partially aerobic soil profile rich in organic matter. This aerobic soil profile is characterised by the accelerated decomposition of peat. This decomposition of peat results in a release of CO2, which has a potentially large effect on the ecosystem’s annual CO2 balance. The annual CO2 balance in a grazed drained peat grassland ecosystem in the Netherlands is determined and the effect of different levels of drainage on this annual CO2 balance is evaluated. It is demonstrated that the restoration of peat grassland ecosystems harbours great potential for CO2 sequestration.

The succession of chapters in the thesis follows a logic in which the temporal and spatial scales of the CO2 exchange processes increase. It navigates from instant grass sward processes to instant ecosystem processes, which then aggregate to annual patterns of CO2 exchange in grassland ecosystems. These annual patterns are subsequently used to explore the potential role of the biosphere in general and the grassland biome in particular in a net CO2 sequestration from the atmosphere.

In Chapter one, the general introduction lays out the processes in the global C cycle at successively smaller time scales, which constitute a complex system of negative feedbacks working towards a stabilisation of the atmospheric CO2 concentration. Since the start of measurements on the atmospheric CO2 concentration a discrepancy has been observed with anthropogenic CO2 emissions. This discrepancy shows a consistent net CO2 sequestration from the atmosphere, although the magnitude of this net CO2 sequestration varies strongly among years. An indication for a role of the biosphere follows from the annual cycle of the atmospheric CO2 concentration, which concurs with the growing season. This annual cycle shows that the atmospheric CO2 concentration sees a sharp drop around spring and gradually increases again afterwards. The relative occurrence of different C isotopes in the atmosphere shows that a particularly large net CO2 sink can be observed in the northern latitudes during the growing season, which supports a clear role for the terrestrial biosphere. Measurements of CO2 exchange in tropical rainforests suggest a net CO2 sequestration which is largely driven by regrowth after prior deforestation. Measurements in temperate forests show a more consistent net CO2 sequestration depending on climate conditions. However, grassland ecosystems can be seen as particularly suited to the sequestration of CO2 because of their physiological characteristics and the nature of their soil organic matter. Measurements of CO2 exchange in grasslands on mineral soils support such a role, even when characterised by large inter-annual variations. Grasslands on drained organic soils generally exhibit a net CO2 release because of the decomposition of organic matter in the aerobic soil profile.

Chapter two starts at the lowest level of process aggregation in this thesis. It introduces an experiment in which the photosynthetic activity in an in vivo grown grass sward was measured under laboratory conditions in summer and autumn. This opens a perspective on the seasonal course of photosynthetic characteristics. Photosynthetic rate was measured as a function of irradiance, temperature (15-30 ºC and 10-25 ºC) and ambient CO2 concentration (200-700 μmol mol-1). The dynamics in the response of photosynthetic activity were analysed by applying a process description where photosynthetic rate is determined by both electron transport rate in the thylakoid membrane and carboxylation at low irradiance. However, it shows that it is difficult to apply these essentially biochemical processes to in vivo leaf area. It proved to be not possible to consistently parametrise the underlying equations. Aggregated hyperbolic response functions instead show that the initial leaf photosynthetic rate at zero irradiance consistently decreased with temperature; the initial rate consistently increased with ambient CO2 concentration. The photosynthetic rate at saturating irradiance was never actually attained and therefore rather was a hypothetical value resulting from the response at lower irradiance. In general, photosynthetic rate appeared to decrease with temperature in summer and show a temperature optimum in autumn.

Chapter three reports on aerodynamic gradient CO2 flux measurements done in a grazed drained peat grassland during a period of two years. The measured instantaneous net CO2 flux was separated into a respiratory and a gross assimilatory CO2 flux. The respiratory CO2 flux responded to temperature in a Q10 type of relationship, whereas the assimilatory CO2 flux primarily responded to irradiance hyperbolically. Low temperature appeared to be strongly limiting the initial response of the assimilatory CO2 flux to irradiance during a substantial part of the growing season. This is a consequence of the concurrence of low irradiance and low temperature in the early periphery of the day. An effect of aerial vapour pressure deficit and calculated surface conductance on the assimilatory CO2 flux could not be detected, possibly as a result of the relatively maritime weather conditions during measurement. A clear correlation was found between the assimilatory CO2 flux at saturating irradiance (a measure for ecosystem photosynthetic capacity) and the respiratory CO2 flux at reference temperature (a measure for metabolically active biomass). The respiratory CO2 flux responded stronger to temperature than the assimilatory CO2 flux. The diurnal cycle of the instant assimilatory and respiratory CO2 fluxes shows that net CO2 sequestration was enhanced under conditions of high primary productivity and moderate temperatures. Moderate temperatures suppress respiratory activity more than assimilatory activity. These are conditions typically found in the higher latitudes.

Chapter four reports on aerodynamic gradient energy flux measurements done in the same grazed drained peat grassland during the same two consecutive years as in Chapter three. It is measured how the downward net irradiance dissipates into an upward latent heat flux (associated with evapotranspiration) and an upward sensible heat flux (conductive heat flux from the surface). The latent heat flux is strongly influenced by the ecosystem’s surface conductance (much derived from its stomatal conductance). Both upward heat fluxes more or less function as communicating vessels. If the latent heat flux is impaired because of low surface conductance, the surface temperature increases. The increased difference between surface and air temperature subsequently increases the sensible heat flux. Their interrelation is reflected in the Bowen ratio, which is the ratio of the sensible heat flux to the latent heat flux. A high Bowen ratio thus points at an impaired latent heat flux and a reduced ecosystem surface conductance. Several months during the growing season were identified with an increased Bowen ratio, thus pointing at a limiting ecosystem surface conductance and thereby suppressing the assimilatory CO2 flux more than the respiratory CO2 flux.

Chapter five reports on eddy correlation CO2 flux measurements done during a growing season in a grazed peat grassland at two different drainage levels. The measured instantaneous net CO2 flux was again separated into a respiratory and a gross assimilatory CO2 flux. The experiment aims to measure the difference in the respiratory CO2 flux between both drainage levels, which is a measure for the difference in decomposition of peat in both aerobic soil profiles. Direct comparison of the respiratory CO2 fluxes was not possible because by far the greatest part of these fluxes are associated with living biomass and not with the decomposition of peat. Relative differences were too small to be detectable. A different approach was thus chosen in which the instantaneous net CO2 flux was treated as a process as such, responding to the gross assimilatory CO2 flux and temperature. The ecosystem net CO2 flux tended towards sequestration at an increasing assimilatory CO2 flux and towards release at increasing temperature. Comparison of the net CO2 fluxes at both drainage levels at zero assimilatory activity and a temperature of 15 ºC indicated that the net CO2 release from the peat grassland with deep drainage was higher by 0.012 mg CO2 m-2 s-1 than the net CO2 release from the peat grassland with a more shallow drainage. At a hypothetically constant temperature of 15 ºC this translates into an annual difference of 375 g CO2 m-2 y-1 or 3.75 ton CO2 ha-1 y-1.

Chapter six aggregates the instantaneous CO2 fluxes and their environmental parameters from Chapters three and five to monthly values. This results in the annual cycle of CO2 exchange at a time resolution where effects of differences in weather conditions on the CO2 flux components and the annual net CO2 exchange can be properly assessed. It shows that the annual cycle of the gross assimilatory CO2 flux responded strongly to irradiance with a baseline response of 2.20 g CO2 MJ-1. The baseline response was mediated by limiting effects of low temperature and high vapour pressure deficit. Low temperature had relatively little impact on the annual assimilatory CO2 flux as it generally concurred with low irradiance. High vapour pressure deficit appeared to be a major limiting factor in the annual assimilatory CO2 flux as it was felt at high irradiance. The annual cycle of the respiratory CO2 flux responded to both temperature and assimilatory CO2 flux, although the effects could not be really separated. The response to temperature followed a Q10 type of relationship, sharply increasing beyond monthly average temperatures of 5 ºC. In its response to the assimilatory CO2 flux the respiratory CO2 flux showed clear hysteresis, with higher values in the 2nd half of the growing season. This can be related both to higher temperatures and to an increase in amount of decomposing senesced organic matter as the season progresses. For the grazed drained peat grassland ecosystem annual net CO2 releases of 623 and 920 g CO2 m-2 y-1 were calculated, which are considered to be a lower boundary for the decomposition of peat in the aerobic soil profile. An export of C in dairy produce which is calculated to be equivalent to approximately 200 g CO2 m-2 y-1 should be added to the ecosystem respiratory CO2 flux depending on the degree to which this C is considered to be respired outside the ecosystem.

In Chapter seven, the general discussion reflects on the research methodology and weaves together the processes at the successive levels of aggregation as explored in Chapters two to six. It demonstrates how patterns of photosynthetic activity in the grass sward re-appear in instant ecosystem CO2 exchange and how patterns of instant ecosystem CO2 exchange translate into an annual cycle of CO2 exchange. In the annual cycle of CO2 exchange higher levels of assimilatory activity lead to a higher net CO2 sequestration whereas higher temperature through its dominating effect on respiratory activity leads to a lower net CO2 sequestration. This emphasises the significance of the climate of an ecoregion, where a moderately warm but productive growing period and low temperatures outside the growing period enhance net CO2 sequestration. Such conditions particularly apply to ecosystems in the temperate and boreal climate zones.

Drained grasslands on organic soils are generally characterised by a net CO2 release originating in the decomposition of organic matter in the aerobic soil profile. In this study an annual net CO2 release of at least 600-900 g CO2 m-2 y-1 was found. It was also shown that drainage depth has a possible effect on the CO2 release. Further calculations on basis of literature indicate that the restoration of drained peat grasslands to undrained conditions could ultimately result in extraordinarily high levels of net CO2 sequestration as a result of new peat formation.

The observed pattern of ecosystem net CO2 exchange in response to assimilatory activity and temperature returns in the annual fluctuations in the biospheric net CO2 sequestration from the atmosphere. Although this net sequestration on average equals 50% of the anthropogenic CO2 emissions, the actual percentage varies among years between 20 and 80%. Closer observation shows that much of this fluctuation concurs with fluctuations in global temperature. Warmer years are characterised by a lower net CO2 sequestration than cooler years. The absolute net CO2 sequestration has been increasing consistently since the start of structured measurement of the atmospheric CO2 concentration. This shows that the negative feedback from the biosphere on atmospheric CO2 has been responding well to an increasing atmospheric CO2 concentration and a longer growing season as a result of rising temperature. It is discussed how grasslands could have a substantial role in this biospheric net CO2 sequestration. It can be argued that the biospheric processes of C capture and C loss in general support a regulating role for the biosphere in between atmospheric CO2 and long-term C sedimentation. Whereas this annual sequestration has a large impact on the course of the atmospheric CO2 concentration, it is of a minute size relative to the soil C content. This makes its direct observation in measurements of soil C content very improbable.

This thesis shows how processes of grassland CO2 exchange result in an annual CO2 balance. It highlights the causes for fluctuations in this CO2 balance. It illustrates how the biosphere as a whole exerts a strong and direct influence on the atmospheric CO2 concentration by instantly sequestering large amounts of CO2. It discusses why grasslands could have a large role in this biospheric CO2 sequestration, even though the Dutch pastureland in this study is in multiple respects uncharacteristic of natural grasslands. It is argued that reinforcing the natural role of the biosphere by restoration of degraded ecosystems such as peatlands is an effective and efficient approach to mitigating anthropogenic CO2 emissions.