Sunset in Yangambi (DRC)


During COBADIM we will focus on eight commercially and ecologically important tree species. These include five species that are ring forming: Milicia excelsa, Etandrophragma angolense, Etandrophragma candollei, Pericopsis elata, Millettia laurentii. These ring forming species are all important timber species exported predominantly towards the EU. Millettia laurentii is commonly used in reafforestation of abandoned farmland, while wild specimens remain listed as endangered by the International Union for the Conservation of Nature (IUCN). The remaining three ecologically important tree species are not ring forming and include: Macaranga spinosa, Scorodophleus zenkeri, Gilbertiodendron dewevrei which represent large fractions of secondary regrowth, mixed and mono-dominant forest types in the central Congo Basin (6.3, 24 and 65% basal area, respectively).

For reference: research methodology as outline in the objectives below is referenced in the progress Ghantt chart.

Objective 1: Quantify long-term observational relationships of tree phenology with climate based on novel observational datasets collected in the Congo Basin

Recently recovered historical observations of tree phenology at Yangambi, located within the central Congo Basin and spanning up to 20 years (~1938-1958) for more than 500 tree species, have the potential to elucidate the sensitivity of central African rainforest phenology to climate variability. In COBADIM I will use the data as gathered through a novel citizen science platform to interpret these phenological records. The phenological records and matching climate data, were gathered by the National Institute for the study for Agronomy in Belgian Congo (INÉAC), with local headquarters in Yangambi. Currently measurements throughout the whole of the Congo Basin are being recovered by dr. Hufkens. These measurements were made in a government research project and adhered to strict protocols (Belgian State Archives files 6077-6078). I leveraged public participation through the citizen science project Jungle Rhythms to transcribe these paper records. Currently, all Jungle Rhythms data has been transcribed by volunteers and preliminary validation analysis have been executed (Fig. 1a). The transcribed phenology data will be linked to the Yangambi climate records previously transcribed (Fig. 1b). Preliminary analysis of the phenology data demonstrates that the citizen science data has a high degree of accuracy (Fig. 1), compared to expert annotations, corroborating previous research. This supports the use of such data within a research context and creates a direct link to public outreach activities. With these datasets I will test the hypothesis that tropical tree phenology (for our species and in general) is linked to rainfall variability.

Fig. 1 a: 8 years (1948-1956) of Citizen Science transcribed phenological observations from the Jungle Rhythms project (black marks) of flowering, fruiting, fruit on the ground and leaf senescence validated against expert transcriptions (red marks) for Millettia laurentii (96% correct); b: A corresponding time series of precipitation at the Yangambi research station.Vertical dashed grey lines mark years

Objective 2: Quantify wood traits and their relationship to phenology and hydraulic properties

Using the eight tree species identified above I will determine a number of key wood traits linked to changes in water availability and hydraulic resistance to cavitation. In particular, I will measure the variation in vessel and fibre frequency, wood densitometry and the frequency and structure of vessel bordered pits. The action will benefit from access to the Tervuren Xylarium (Xylarium from hereon forth) that maintains a large collection of African tropical wood specimens at the Royal Museum for Central Africa (RMCA, Belgium). The Xylarium collection and research staff there will provide training, databases and facilities that can be used to determine these relevant wood traits such as X-ray diffraction, visual and electron microscope imaging techniques. In addition I will be trained how to use the ROXAS imaging software to measure these traits. This anatomical work, and collection of wood cores from stem discs (see Objective 3 below), will be executed during a short research visit under the supervision of Dr. Hans Beeckman, curator of the Xylarium. I will use historical phenology and climatology data together with wood traits and existing leaf traits data in a multivariate analysis to differentiate (anatomical) strategies of tropical tree species to respond to and survive periodical (severe) drought periods.

Objective 3: Develop a multi-proxy tree-ring approach to understand tropical tree drought resilience and growth

Previous studies on Entandophragma angolense and Pericopsis elata have shown ring width and growth are correlated with late season (Sep-Nov) precipitation amount and variability in El Nino strength4. However, many tropical tree species do not form rings and so far it is not clear how the growth of these species varies with climate drivers and how vulnerable they are to drought. To overcome this problem, inter-annual and intra-annual high resolution carbon (d13C) and oxygen (d18O) stable isotope measurements often reveal the seasonal dynamics of rainfall inputs in temperate (Fig. 2) and tropical tree species and can be modelled reliably from meteorological forcing by the host institution (Fig. 2). Thus for all eight species described above, Xylarium archived stem discs from the central Congo Basin will be used to construct high resolution d13C and d18O time series using the Leavitt & Danzer method for holo-cellulose extraction.

However, stable isotope analysis can be costly and labour intensive. Thus this project will also develop novel X-ray microprobe synchrotron records of calcium variations from the same archived stem discs at the SOLEIL Synchotron facility in Paris. In particular, COBADIM will test the hypothesis that the inter- and intra-annual variability of [Ca] in tree rings is mechanistically linked to evapotranspiration rates of trees. To test this hypothesis I will conduct a climate control experiment four species. In particular I will focus on two African tropical tree species, one with rings (Pericopsis elata) and another without rings (Macaranga spinosa) and two temperate tree species (Quercus robur / Pinus pinaster) that also form annual rings, with four replicates each. All tree species will be subjected to four variations in relative humidity (Rh) and volumetric soil water content (VWC) manipulations using isotopically labelled soil water (Fig. 3). A large number of individuals (40) of each species are being grown from propagation specimens at the Botanical Garden Meise (Belgium) by Dr. Reynders (manager of the living plant collection). Over the growing season gas exchange measurements will be made on these plants to obtain information on rates of evapotranspiration and stable isotope evaporative enrichment. These plants will then be destructively harvested and the annual rings (defined by cambial marking if necessary) measured at high-resolution for [Ca] and stable isotope composition. In addition, stem samples from each of the species will be measured for cavitation resistance on the Xyloforest platform at the University of Bordeaux using the method of Cochard et al. (2002). The output of the manipulation experiment will consist of the responses in [Ca], d13C and d18O and evapotranspiration rates and their relation to variations in both relative humidity and volumetric soil water content. An example of such output from a previous analysis of stem samples is provided in Figure 2 (d13C and d18O, only). We will use these data to test the hypothesis that Ca deposition in annual rings is linked to variations in transpiration rate. In COBADIM we will develop the necessary theory to link [Ca] variations to changes in transpiration rate and climate.

Objective 4: Develop a multi-tracer model to predict the functional response of central African species to future climate scenarios

The MuSICA model developed by the host institution provides the perfect framework to tie together all the findings from Objectives. MuSICA predicts changes in evapotranspiration and carbon uptake in response to rapid and long-term variations in climate4,5, plant water potential6. More so, MuSICA predicts the isotopic fractionation of CO2 and water in ecosystems at high temporal resolution7,8 and how all these processes drive the d13C and d18O signals in tree rings (Fig. 2). Using the theory developed in Objective 3, COBADIM we will implement this new relationship between [Ca] and transpiration in MuSICA. For the final part of this project MuSICA will be fully parameterised for the different tropical tree species (gas exchange, cavitation resistance and phenology) and for typical site conditions characterised in Yangambi. Using the local meteorological data (Fig. 1b) MuSICA simulations will be made for a variety of past, current and forward-looking representative concentration pathway (RCP) climate scenarios. The performance of the model will then be validated against the [Ca], d13C and d18O signals in tree-rings for the eight species obtained in Objective 3. The model can explore the effects of tree competition on transpiration and CO2 sequestration and can thus help formulate new hypotheses on how different combinations of our studied species can affect the carbon sequestration potential of African tropical forests. This is especially important as recent studies suggest most CO2 uptake is dominated by a few key species in tropical forests. Thus we will test this hypothesis using MuSICA simulations and explore how drought will affect the C sequestration potential of key (commercially important) tropical tree species of the central Congo Basin.

Fig. 2 - Long-term high resolution measured and modelled wood core stable isotope values (C and O) for Pinus pinaster; VPDB, Vienna Peedee belemnite; VSMOW, Vienna Standard Mean Ocean Water
Fig 3. – The different experimental relative humidity (Rh) and volumetric soil water content (VWC) treatments.