Predicting Impacts and Communicating Forecast Uncertainty in the Case of Drought Dr. Cody L. Knutson National Drought Mitigation Center University of Nebraska-Lincoln, USA
Predicting Drought Impacts through Understanding http://droughtreporter.unl.edu/
Drought Impact Reporter Public Input
Drought Impact Reporter database to help gain a broader understanding of the full range of potential drought impacts by sector Areas of Future Research develop standardized impact assessment methodologies currently investigating modeling strategies to better estimate the economic impacts of agricultural drought losses need common assessment methodologies in other sectors develop standardized methodologies for qualitative and quantitative analyses of the effects of drought at the national, state, and local levels Use information and models to predict potential impacts at various scales and drought severity levels depending on forecasts
Advanced Hydrologic Prediction Service Flood early warning system Network of 3568 stream gauges Providing current river flow/stage data Up to 7-day forecasts at many forecast points http://www.weather.gov/ahps/ 90-day probabilistic forecasts at 1,376 forecast points
NOAA/NWS and NDMC Low Flow Studies Surveys to identify potential low flow impacts in the Upper Mississippi, Upper Missouri, North Platte, Trinity, and Red River basins for inclusion in AHPS
River Flow and Drought Impact Information Mississippi River near Anoka 2120 (cfs) 2000 2000 1200 1000 750 If this flow is sustained for 72 hours, surface water appropriations in all contributing upstream major watershed may be subject to suspension Hydropower facilities at Coon Rapids Dam, St. Anthony Upper Lock and Dam, St. Anthony Lower Lock and Dam, and Lock and Dam #1 will implement measures to adjust river flows as part of the Mississippi River System-Wide Low-Flow Management Plan. May trigger a state Drought Watch as described in the MN Department of Natural Resources Drought Response Plan. St. Paul Regional Water Services begins voluntary conservation plan St. Paul Regional Water Services begins sprinkling restrictions St. Paul Regional Water Services implements mandatory conservation (Photo taken by Ken Dewey, 2001)
Future Work for River Forecast Centers North Central River Forecast Center Figure 3. An example of a seven-day river forecast hydrograph with indicators of critical low flow and flood levels. The red line indicates the flood level. The tan line indicates the critical low flow level at this forecast point. The blue line shows the current river level and the green line is the actual forecast.
Communicating Forecast Uncertainty
NOAA OGP funded research among agricultural producers in Nebraska When the short-term forecast really comes into affect in my life if they say 40% chance, I might change what I m going to do. If they re saying 80-90% chance of rain for the next day, and it feels like it s going to rain, then I ll really pay attention to it (Nebraska rancher 2005) I would think as much data as they have available they ought to be up there around the 80% or 90%. Don t give me this 20% chance of rain today or 60%. I m talking about 85% or 90%. I don t see why they can t, but they re not. It s terrible (Nebraska rancher 2005)
Weather and Climate Information Sources NOAA - National Weather Service DTN Where do you get your weather and climate information? Radio Television (excluding Weather Channel) The Weather Channel (TV and online) Internet (excluding Weather Channel) Newspapers, newsletters and magazines US Drought Monitor University of Nebraska God, neighbors, private forecasters NASS, NOVA, folklore, joint pain, Cattle Fax, clairvoyant, animal behavior, own forecast Radio Television Internet Respondents 26 22 17 13 13 9 7 6 3 2 each 1 each
The influence of weather information and forecasts on my decisions is limited by? The influence of weather information and forecasts on my decisions is limited by: Current/ Recent Past Conditions Short-term Forecasts Long-term Forecasts N Mean N Mean N Mean The accuracy of the weather information and forecasts 83 2.69 a 80 2.85 a 81 3.25 The reliability of the source of weather information and forecasts 80 2.56 78 2.91 78 3.03 My ability to apply the weather information and forecasts 80 1.95 a 78 2.08 ab 77 2.23 b Availability of weather information and forecasts for this area 81 1.95 79 1.95 78 1.96 My financial situation 83 2.10 79 1.76 80 1.98 Timeliness of weather information 78 1.87 76 2.00 75 1.9 My understanding of weather information 80 1.91 b 76 1.76 a 73 1.84 ab Views of others (e.g., banker, friends) 82 1.48 76 1.40 77 1.52
Influence of weather and forecasts information on decision making: Target Analysis Influenced this decision in 2004: Current and Recent Past Conditions Short- Term Forecasts Long-Term Forecasts N Mean N Mean N Mean Summer livestock decisions (e.g., stocking, culling) 76 3.6 b 76 2.4 a 76 2.9 a Summer cropping decisions (e.g., pesticides, fertilizer, watering) 64 3.3 b 67 3.2 b 65 2.1 a Crop/livestock marketing 75 3.3 b 70 2.4 a 70 2.6 a Fall livestock decisions like stocking and feeding 74 3.2 b 68 2.3 a 69 2.6 a Spring livestock decisions such as stocking options 77 3.2 c 74 2.0 a 77 2.5 b Harvest and post-harvest decisions like fall tillage 39 2.9 b 41 2.6 ab 40 2.1 a Spring cropping decisions (e.g., crop type, seed variety, tillage) 50 2.7 b 48 2.2 a 50 2.4 ab
The long terms I look at but I don t put much salt into it. I think they re just wasting their time on those over 10 days. Sure, they have 90-day forecasts but that s just complete nonsense. No, I look at them but they don t mean nothing to me. Now, the 3-4 day forecasts, I use them a lot, especially in the wintertime; big time in the wintertime. You want to know if there s going to be a storm or something coming to get the cattle in behind shelterbelts or move them to where there s protection. Summertime, you just look at it 2 or 3 days out (Nebraska Rancher 2005).
Future Research in Communicating Uncertainty Investigate the use of specific forecasts for specific tasks - detailed target analysis (summer cropping decisions; NOAA short term forecasts such as 1-5 day QPF and 6-10 day outlook; US Seasonal Drought Outlook) Investigate and foster accessibility to local forecast information channel preferences Investigate and develop relevant tools to meet user needs - decision support systems that portray uncertainty Investigate and foster educational and participatory opportunities related to climate and weather, including communicating forecast uncertainty building understanding, trust, and buy-in
N I D I S A dynamic and accessible drought information system that provides users with the ability to determine potential drought impacts and associated risks and the decision support tools needed to better prepare for and mitigate the effects of drought.