SE CSC Science in the US Caribbean Adam Terando, USGS SECSC - PowerPoint PPT Presentation

SE CSC Science in the US Caribbean Adam Terando, USGS – SECSC Climate models, frog calls, and the path towards long-term adap6ve species management

With special thanks to: Jaime Collazo, NC Coop Fish and Wildlife Research Unit Jared Bowden, NCSU, Applied Ecology

Guajataca Dam, Quebradillas, PR. Source: The Atlan6c

Utuado, PR. Source: NY Times

Corozal, PR. Source: The Atlan6c

Yabucoa, PR. Source: The Atlan6c

San Juan, PR. Source: The Atlan6c

Toa Alta, PR. Source: The Atlan6c

Toa Baja, PR. Source: The Atlan6c

Naranjito, PR. Source: The Atlan6c

Puerto Rican Parrot ( Amazona vi*ata )

MOTIVATION Chadwick, R. 2016. Sub-tropical drying explained. Nat. Clim. Change .

25 species Endangered PR Crested Toad 17 Eleutherodactylus • 2 endangered • 14 at risk Amphibians in Puerto Rico

How will subtropical drying affect amphibians on the island? El Yunque Rainforest

How will subtropical drying affect amphibians on the island? Guánica Dry Forest

How will subtropical drying affect amphibians on the island?

BROADER CONCEPTUALIZATION Wise et al. (2014) Global Environmental Change.

VULNERABILITY How wide is this space? FORCING What is it’s trajectory?

Ul6mately, trying to evaluate candidate strategies for adap6ve management • Passive management in marginal habitats • Translocate Popula6ons • Habitat acquisi6on

18°N Khalyani et al. (2016)

Exposure/Response Func6ons Time between rainfall events Risk of Extinction Guanica (dry forest) Maricao (wet forest) Present 2060 Exposure Exposure Time

Exposure/Response Func6ons Rates Egg Development/Hatch Rates Risk of Extinction Risks Present 2060 Exposure Exposure Guanica

Ground heat flux Cloud-based height April Rainfall > Soil moisture 9mm/day climate-response func:on

CLIMATE MODELING FIELD ECOLOGY

Expect Sub-tropical Drying in This Region Chadwick, R. 2016. Sub-tropical drying explained. Nat. Clim. Change .

Global Climate Models are s6ll very coarse

Exposure/Response Func6ons Time between rainfall events Risk of Extinction Guanica (dry forest) Maricao (wet forest) Present 2060 Exposure Exposure Time

Insights from Downscaling Time between rainfall events Risk of Extinction Guanica (dry forest) Maricao (wet forest) Present 2060 Exposure Exposure Time

1) Projec6ons that reflect reality given constraints of GCMs and oceanic context. 2) Simulate precipita6on and other covariates response to the anthropogenic forcings across Puerto Rico. **Elicit expert knowledge to select relevant climate variables .

Chose to use dynamical downscaling To 30-km To 10-km To 2-km

OUR GOAL: 2-KM Horizontal ResoluWon With Hourly Output To 30-km Using mulWple RCM-GCM combinaWons To 10-km To 2-km

Weather Research and ForecasWng Model (WRF) Regional RSM Spectral Model NHM (RSM) and the Non-HydrostaWc Model (NHM)

Collabora6on with Vasu Misra Weather Research and ForecasWng at FSU Model (WRF) Regional RSM Spectral Model NHM (RSM) and the Non-HydrostaWc Model (NHM)

Select Global Climate Models to Downscale Scenario RCP8.5 (High GHG Emissions) Historical (1986-2005) and Future (2041-2060) * indicates completed GFDL-CM5 CNRM-CM5 CCSM4 RSM-NHM WRF WRF-CCSM4* RSM-NHM-CCSM4* WRF-CNRM-CM5* RSM-NHM-GFDL-CM5

Experimental Design for Regional Climate Modeling • THREE GCMs – CCSM4, CNRM5, GFDL-CM3 • TWO RCMs – WRF, NHM-RSM • TWO 20 year periods – 1986-2005 (past) – 2040-2060 (future) – RCP 8.5 – high fossil fuel emissions scenario 42

Many More Physical Variables Available (and relaWonships between variables are maintained) • Surface – Rainfall, Temperature, Humidity, winds, soil moisture/ temperature, runoff, evapotranspira6on, pressure • Above canopy – As above, plus others – Mixing height, ver6cal winds • Radia6on – Incoming, outgoing, diffuse, net, cloud frac6on • Diagnos6c Variables – Height of cloud base, – Sta6s6cal : Heat Wave dura6on, extremes, percen6les, etc.

Many More Physical Variables Available • Surface – Rainfall, Temperature, Humidity, winds, soil moisture/ temperature, runoff, evapotranspira6on, pressure Time, Storage, and Processing • Above canopy Constraints => Cannot Retain All – As above, plus others – Mixing height, ver6cal winds Variables at All Time Steps • Radia6on – Incoming, outgoing, diffuse, net, cloud frac6on • Diagnos6c Variables – Height of cloud base, – Sta6s6cal : Heat Wave dura6on, extremes, percen6les, etc.

2-Day Stakeholder workshop hosted by CLCC in San Juan to refine climate model output

IDEA IS TO HAVE CLIMATE PROJECTIONS THAT ARE SPECIFIC TO THE DECISION, BUT ALSO RELEVANT TO OTHER SCIENTIFIC/ ECOLOGICAL QUESTIONS

How could climate change affect shade coffee producWon? Providing public goods

Follow-up workshop in August 2016 to discuss available modeling outputs Providing public goods

Rank climate variables based on ecological significance Used this dialogue to help retain necessary climate model data

Downscaled Climate Variables

Exceeded 1 million CPU hours to accomplish the downscaling for just one of the regional climate models. We reduced ~1 Petabyte of model output to < 20TB with the knowledge of climate variables to retain from prior workshop

Maximum 2-m Temperature Change annual average

PrecipitaCon Change percent change for the annual total

Hourly rainfall bin % difference > 1”/hr ECOREGION ANALYSIS (Subtropical wet forest - wet season)

Projected Changes Soil Moisture

Low-level Cloud FracWon

p(Occupancy | Temperature) p(Occupancy | Temperature) 0.10 1 Local Occupancy Temperature Probability (Psi) 0.15 E.wightmanae 0.9 0.08 0.8 0.10 0.06 Density Density 0.7 0.04 0.05 0.6 0.02 0.5 0.00 0.00 0.4 20 22 24 26 28 30 20 25 30 Temperature (°C) 0.3 Temperature (°C) p(Occupancy | Precipitation) 0.2 p(Occupancy | Precipitation) 5e − 04 0.1 0.0012 4e − 04 0 Precipita6on 0.0010 0 100 200 300 400 500 600 700 800 900 1000 0.0008 3e − 04 Density Density 0.0006 2e − 04 1 0.0004 1e − 04 E.brifoni 0.9 Local Occupancy Proability 0.0002 0.8 0e+00 0.0000 0.7 0 500 1000 1500 2000 − 500 0 500 1000 1500 2000 2500 Annual Precip (mm/yr) 0.6 Annual Precip (mm/yr) p(Occupancy | Dry Seas Soil Moisture) 0.5 p(Occupancy | Dry Seas Soil Moisture) 2.5 0.010 (Psi) 0.4 0.3 Soil Moisture 2.0 0.008 0.2 0.006 1.5 Density 0.1 Density 0.004 0 1.0 0 100 200 300 400 500 600 700 800 900 1000 0.002 0.5 0.000 0.0 ElevaWon (m), PrecipitaWon 0 20 40 60 80 100 0.2 0.4 0.6 0.8 1.0 Soil Moisture (%) Soil Moisture (%) What are the environmental limits of these species?

Use acous6c recorders to es6mate occupancy of three species across environmental gradients

EsWmate occupancy based on recorded calls

Precipita6on 1 Local Occupancy Probability La6tude E.wightmanae 0.9 0.8 0.7 0.6 (Psi) 0.5 0.4 Longitude 0.3 How could these 0.2 0.1 gradients change 0 0 100 200 300 400 500 600 700 800 900 1000 with climate 1 E.brifoni 0.9 Local Occupancy Proability 0.8 change? 0.7 0.6 0.5 (Psi) 0.4 0.3 0.2 0.1 0 0 100 200 300 400 500 600 700 800 900 1000 ElevaWon (m), PrecipitaWon

NEXT STEPS

El Yunque Caribbean Na6onal Rainforest Next steps: Explore resilience of windward slopes

El Yunque Caribbean Na6onal Rainforest PotenCal to couple to WRF-Hydro Model

Hybrid downscaling

Select Global Climate Models to Downscale Scenario RCP8.5 (High GHG Emissions) Historical (1986-2005) and Future (2041-2060) * indicates completed GFDL-CM5 CNRM-CM5 CCSM4 RSM-NHM WRF WRF-CCSM4* RSM-NHM-CCSM4* WRF-CNRM-CM5* RSM-NHM-GFDL-CM5

Global Climate Models to Downscale Scenario RCP8.5 (High GHG Emissions) Historical (1986-2005) and Future (2041-2060) CNRM-CM5 GFDL-CM5 CCSM4 RSM-NHM WRF WRF-CCSM4 RSM-NHM-CCSM4 WRF-CNRM-CM5 RSM-NHM-GFDL-CM5 ARRM-WRF-CCSM4 ARRM-WRF-CNRM-CM5

CCSM4 (GCM) OBS Combining sta6s6cal and dynamical downscaling approaches

OBS RCM Combining sta6s6cal and dynamical downscaling approaches

Sta6s6cal Model OBS Combining sta6s6cal and dynamical downscaling approaches

Hybrid OBS Combining sta6s6cal and dynamical downscaling approaches

Taking occupancy modeling a step further. Is reproduc6on occurring at occupied sites? Are sites being occupied by a few individuals or by “many”? Plus gene6c work to establish popula6on structure.

Augment field work with terraria experiments to test eco- physiological limits (w/colleagues at Univ. Puerto Rico)

Geo Data Portal (GDP) Web-based access to and processing of global change data to address climate and landscape change

THANKS! QUESTIONS?