FOOD SECURITY AND AGRICULTURE IN THE ANDES
Climate impact indicators
Global essential climate variables/indicators used
ECMWF seasonal forecast System 5 (hindcast period 1993-2015) for precipitation and mean, maximum, minimum temperature for all initializations (7 lead times).
Pre-processing: S5 daily data will be aggregated to monthly and quarterly data for downscaling
Regional/local indicators used and produced
Probabilities of rainfall and temperature (terciles)
Probability of exceeding certain threshold
Probability of having a season similar to an analogue year
Number of days with rain (without)
Global data to regional/local scale
Step 1: S5 data download: ECMWF seasonal forecast System 5 (S5, hindcast period 1993-2015) for precipitation and mean, maximum, minimum temperature was provided by SMHI for all initializations (7 lead times). S5 is retrievable from the CDS catalogue.
Step 2: Bias adjustment, downscaling, and quality control: Bias adjustment procedures were applied to S5 data outputs. Quality control analysis was applied to stations data. Data were checked for outliers, duplicates, and inconsistency. S5 monthly rainfall outputs were further downscaled to 0.05 degrees gridded data for all pilot basins. Downscaling was performed using the IRI´s Climate Predictability Tool (S5 rainfall as predictor, CHIRPS data as response variable) and tailored for sectoral applications.
Step 3: Station data processing: Climate indices with stations data (reference climatology) were calculated (consecutive dry days, consecutive wet days, yearly total precipitation amount, 95thprecipitation percentile, heat waves or warm days and cold waves or cold nights). All climate indices for the period 1981-2010 were calculated with stations data.
Step 4: Presentation of downscaled outputs: Downscaled outputs are presented in term of probability of exceedance of upper and lower percentiles, probability of exceedance of certain critical threshold, probability of having similar conditions as in El Niño or La Niña years, among others.
Step 5: Web mapping: Tailored downscaled outputs and vulnerability maps were uploaded into a web system. The system was developed using THREDDS data server technology in order to enhance the interaction of users with the demonstration products. This allows users to visualize climate threats and the vulnerability of the basins, in order to facilitate decision making based on high quality and high resolution data.
Step 6:Feedback with NMHS and users: Meeting with NMHS´s technicians to discuss the concept of the study case took place. A communication (provider-user-provider) strategy with local authorities was developed. The strategy considers both ways of communication from providers to users, and feedback from users to providers. Online workshops with local authorities to explain how to interpret and use the information provided, including limitations, took place.