The multi-year impacts of the 2015/2016 El Nino on the carbon cycle of tropical forests

A key challenge in Earth System science is to understand the role the biosphere plays in ameliorating or accelerating global atmospheric change. The global carbon cycle is of particular interest: the terrestrial biosphere appears to act normally as a sink for atmospheric carbon dioxide, but under extreme conditions, in particular during El Niño events, appears to switch from being a sink to a net source of carbon dioxide to the atmosphere. The tropical land biosphere in particular dominates the interannual variability of the carbon cycle, and this variability appears to be increasing over recent decades. This suggests that the tropical land biosphere may be approaching some critical transition, that may result in it becoming an accelerator of climate change.

While the sensitivity of the tropical carbon cycle to climate perturbations has frequently been modelled and more recently estimated from atmospheric mass budgets and remote sensing, there has been almost no direct field validation of the impact of the El Nino on the tropical carbon cycle. The Earth is currently (2015/2016) experiencing its strongest El Niño event for many decades, with unusually high temperatures across the tropics and drought in many parts of the tropical Americas, Africa and Asia. Unlike the case during previous strong El Niños, we now have established a global network of detailed carbon cycle process studies across the tropics, supported by a patchwork of short-term funding.

Such an opportunity may not come again for decades, and we are well-placed to seize it. Our plots monitor forest carbon cycling at seasonal and monthly timescales, and span across the tropics (in Belize, Bolivia, Peru, Brazil, Ghana, Gabom, Ethiopia, and Malaysia). They also span gradients of human modification ranging from old growth through different intensities of logging, fire occurrence, agroforestry and other natural disturbance including hurricanes. This provides a unique opportunity to attain a direct empirical understanding of how carbon cycle processes change over and after an El Nino, including productivity, allocation, autotrophic and heterotrophic respiration. We propose to monitor 80 plots across the tropics through this El Nino and beyond until end 2017, at monthly or seasonal resolution.

We are uniquely placed to closely track the carbon cycle of forests through this event and for several years after, potentially providing a definitive answer to how much tropical forest carbon cycle responses contribute to the interannual variation of the biosphere carbon sink, and which component processes determine this interannual variation. This information will be use to validate and improve land biosphere models, including the JULES land surface model that sits within the UK Met Office Earth System Model.

In this proposal we see to utilise this network to answer several key questions:
1. What are the local immediate and multi-year carbon cycle responses to drought, and how consistent are these across the tropics?
2. How does a ~1 degree C rise in temperature affect carbon cycle processes across gradients of drought and disturbance?
3. Which components of productivity and respiration dominate interannual variability, and what can this tell us about the correspondence between short-term variability and long-term resilience?
4. Is there evidence that forest productivity (NPP) is maintained through the event via a decline in non-structural carbohydrate reserves?
5. How do human-modification (logging, tree crops, past fires) and natural disturbance (hurricanes, elephants) affect carbon cycle sensitivity to the El Niño?
6. What are the impacts of fire on the carbon cycle of our forest plots in Brazil, that have burned in this El Niño?

This study would provide the clearest picture yet available of the main mechanisms of carbon cycle perturbation within intact tropical biomes, information that will be used to directly inform and test carbon cycle sensitivities in global biosphere models.

In addition to its academic and scientific impact, the work proposed will further build scientific capacity in partners in eight tropical countries (Belize, Bolivia, Brazil, Peru, Ethiopia, Gabon, Ghana, Malaysia), training field technicians and involving scientific partners in paper development.

We anticipate that the outcomes of this work will directly inform land biosphere models. Project Partner Peter Cox will lead a close model-data comparison using the JULES land surface model, likely resulting improvements and insights will be incorporated into versions of JULES used in the next generation UK Met Office Earth System Model. We will work with modelling groups to establish runs of models such as JULES and other models driven by our observed data, so that model-data comparisons can be presented at the workshop. Project Partners Peter Cox and Nate McDowall will assist in establishing model runs to compare with our data, and ensuring good connectivity with the modelling community.

We will continue to disseminate reports of our field research "from the front line", including using videos, blogs and social media

To maximise impact, we are proposing to hold a synthesis workshop (~50 attendees) on El Niño impacts on tropical forests. This will focus both on the scientific community (both field researchers but also modellers and atmospheric scientists), but will include an important public outreach dimension. We will also invite key science journalists to participate and fully or partially attend.

The aim of this conference will be to:
(i) present and discuss results from this grant proposal, including descriptions of individual sites and syntheses across sites. All project partners would be invited, funded by this grant.
(ii) include other researchers (in the UK and beyond) to present results. This would include other field based research.
(iii) include a strong representation from the land biosphere modelling community, with the explicit aim of comparing model simulations with observations, and seeking to identify areas of model improvement and development.
(iv) include a public component of the workshop where headline results will be presented to the general public and media. All presentations will be recorded as podcasts with attached slides to enable students and researchers in tropical countries to follow (something we have done for several previous conferences). The conference will hold two keynote public lectures on consecutive evenings.
(v) produce a number of papers (e.g. analyses along particular gradients or particularly interesting sites) aimed for a special issue of a journal, as well as one or more high impact synthesis papers aimed at Science or Nature. We will also work with participating science journalists (and other journalists) to produce articles for the general public.