Early warning System for critical transitions in Amazonia : Framework

1. Early warning System for critical transitions in Amazonia : Framework

2. Coupled Atmosphere-Land surface Analyse and new ESM results DGVMs test urgent issues rivers Coupled LUC Critical Ecosystem Services People Scenario analysis Policy analysis Land-use change model Blue-print Early-Warning System

3. Develop a proposal (‘Blue-print’) for an early warning system for imminent tipping points

4. Early Warning System What to warn about? • Not only critical transitions, but also gradual change? • Which changes are the biggest concerns of stakeholders? • Degradation of ecosystem services in what time scales (years – decades) ? • How do we translate large-scale change into local impact? • Are there likely policy options in response?

5. Amazonia maintains its own regional water cycle

6. Amazonia stores CO2 in biomass On area basis not extremely high, but there is so much of it: total 86 Pg C Globally in atmosphere: 760 Pg

7. High(est?) Biodiversity

8. People!

9. What is needed to preserve ecosystem services? Does most of it come down to preserving forests?

10. How to define a threshold/change • Magnitude of impact • Time and spatial scale • Threshold in impact or in forcing?

11. How to define thresholds Good et al, 2012

12. What kind of thresholds can we warn for? Good et al, in prep Dry season length Committed climate change after emissions are stopped (lagged ocean heat uptake) + Emissions set to zero here Forest unsustainable Forest sustainable ‘Regional forcings threshold’ Temperature • IMPACT threshold: Critical change occurs • REGIONAL threshold: Critical change is unavoidable • MITIGATION threshold: Existing policies are not sufficient any more • to avoid critical change

13. Indicators • What would change look like • models, proxy observations • How would the ecosystem approach the change • Variability indicators? • Examples... • Mitigation threshold: • model forecasts of future T , seasonality, land-use • Regional threshold: monitoring and model • T, Precip, land-use • Impact threshold: monitoring • NPP, soil carbon, flammability, hydrology

14. Forecast climate-induced changes to the Amazon till 2100 ‘ 40% deforestation and/or 3 degrees warming leads to savannisation’ (Nobre et al., various)

15. Precipitation change RCP4.5 - 2050 No fire and no LU + fireand LU

16. Dry season length RCP4.5 - 2050 No fire and no LU + fireand LU

17. Biomass RCP4.5 - 2050 No fire and no LU + fireand LU Sampaio et al, in prep

18. Tipping points in Amazonia Tropical forest Savanna state triggered by climate change or deforestation Stability Stability of savanna enhanced by increased droughts and fires Environmental variation Cardoso andBorma, 2010

19. Analysing tipping points with simple models Example: Simple Amazon with variable climate circulation (Meesters, in prep)

20. Would early warning signs be detectable?? • Increased variance swamped in overall variance increase • Time scale of fluctuation short relative to time scale of change itself • No fluctuation signal at all

21. Critical indicators • The basis of such a system is long-term monitoring of critical indicators • These indicators should be quantities that are relatively accessible, and easy to monitor at high temporal and/or spatial resolution. • should represent the variability of the Amazon ecosystem services and other important tipping phenomena => their behaviour near critical transitions should reliably point to imminent change in the state of that particular ecosystem service.

22. Monitoring system • WHICH properties of the Amazon would be important to monitor as indicators? • Existing monitoring • Invest in key new ones • HOW would we detect imminent change from these? • Advanced statistical techniques looking at variability and ‘slowing down’ • Analyse trends as well and compare with models to extrapolate • Analyse model output as guidance

23. EWS framework Remote sensing monitoring system Modeling System In situ monitoring stations • Monitoring system • Modeling System • Analysis Tools • Communications EARLY WARNING SYSTEM

24. Which institution can do this? • Aim: • Manage critical monitoring systems • Re-analyse and re-run coupled forecasts • issue regular ‘state of the Amazon’ reports • CEMADEN – Brazil centre for monitoring and forecasting of natural disasters • Federal institute • Runs several intensive monitoring systems • Now short-term forecasting • Could add long-range forecasting branch • An independent NGO? • More difficult to obtain information and run monitoring • Could more easily advise against government policy

25. Products for this week? • Outline / draft of ‘proposal for an EWS for Amazonia’ • Outline for high-level paper on EWS: the science behind it • An explicit list of monitoring elements • An explicit list of model tools needed

26. Early warning system – draft outline • Components: • Critical ecosystem services (WP1) • Requirements from policy/stakeholders (WP4) • Review risk of collapse (WP3) • Framework for threshold typology (theory) • Review of likely thresholds/ behaviour at threshold crossing (so far, from literature) • Indicators (requirements and listing – WP2,3) • Monitoring options/recommendations (WP5) • Analysis tools and forecast models (WP5, WP3) • Institutional embedding (WP5) • Communication means (WP5, WP4, WP6)

27. Implementation EWS Likelihood collapse Types of thresholds Review thresholds Critical Services Policy need Indicators Monitoring Analysis/forecast Measures (adapt, mitigate) Institutional Communication

28. Format • Policy report • Submittable proposal for tender • Paper on concept of EWS • Review of most important thresholds depending on time and spatial scale • Joint with WP3/WP2, explore the model results (BESM, IPSL, CMIP3, DGVMs?) – update Nobre&Borma (3 degrees-40%), Good et al (reg thresholds)... • Review socio-economic thresholds / indicators (need WP4 to think about this!) • Review on (new) observational and analysis methods for EW? • Only if we can do a TEST. So far not enough data/no tipping points? • Lack of power to detect tipping points (Antoon Meesters) • ..?

30. List of possible variables to monitor • Sea Surface Temperature (SST) - indicator of global-scale change • Precipitation (patterns, quantity, dry season length…) primary driver as well as an ecosystem service that can be affected • Climate modes (ENSO, Atlantic Oscillations, etc) - often correlated indicators of high-impact changes or episodes in Amazonia • River flow and discharge • Evapotranspiration - prime driver of recycling

31. List of possible variables to monitor • overall vegetation productivity changes – [CO2] over the tropical belt + anthropogenic emissions • Monitor a number of permanent plots. Among others, look at PHENOLOGY. • Biomass - remote sensing (eg S-band Radar) and well-referenced growth bands in forest plots across the basin • Water use efficiency from tree-ring & gas exchange monitoring • Remote sensing indices (NDVI , EVI)

32. List of possible variables to monitor • Fires (remote sensing and in-situ observations) – not simply occurrence or area, but also fire effects (e.g. type of vegetation affected and recovery of previously burned areas • Economic indicators, such as the GDP of the region, transport, trade and migration patterns • Exposure and Vulnerability (?)