Extended Abstract (1004K)6B.5
Development of a smart search system for climate services
Mark S. Brooks, North Carolina State Univ., Raleigh, NC; and B. E. Aldridge
The value of timely weather and climate information to some users, especially growers, is undeniably important. The State Climate Office of North Carolina (SCO) at NC State University works closely with researchers and extension scientists to develop agricultural decision support tools. Data for these tools comes from the Climate Retrieval and Observations Network of the Southeast (CRONOS), a digital database of high-resolution surface weather observations. Data and decision support tools are currently disseminated via the web, e-mail, and telephone interaction with citizens. However, many people in rural parts of North Carolina do not have Internet access but could benefit from climate services.
This need compelled the development of a text messaging service that will enable anyone with a text messaging capable mobile phone to interface with CRONOS and basic climate products. We call it CRONOS Smart Search. The program is also accessible on the Internet where users can enter their data request in a single text box. Table 1 illustrates a sample of inputs from users and outputs returned.
Message received from user | CRONOS Smart Search response |
Blackberry chill at Lewiston | Lewiston , NC (LEWS) has 811 blackberry chill units as of Feb 6, 2008, 4:00pm. |
Peanut hours at Clinton | Clinton , NC (CLIN) has 21 favorable leaf spot hours, 33 favorable Sclerotinia hours. Spraying is not yet advised. |
Weather at KRDU on September 5, 2002 at 8pm | Raleigh , NC (KRDU) weather on Sept 5, 2002 at 8pm: temperature=67 F, dewpoint=60 F, winds=3.4 mph from the NNE, no precip, hazy. |
Weather at 1005 Capability Drive , Raleigh , nc on April 1, 2004 | Raleigh , NC ( LAKE ) weather on Apr 1, 1999: high/low: 58/43 F, 0.12 in. precip |
Weather in Salisbury , NC on November 22 2003 | Salisbury , NC (SALI) weather on Nov 22, 2003: high/low: 76/33, no precip. |
help | Help dialogue with examples of how to use the service. |
Table 1 - Example inputs and outputs of the smart search
A nominal amount of language processing was required to parse the user's request and formalize the logic to craft a response. When a message is received, software parses and validates the message, then responds appropriately. Interaction via the webpage asks the user for feedback about how well the information matched their question. Feedback for each request via the webpage is being recorded for future analysis and improvement.
The impact of this tool, quite literally, is putting climate services and CRONOS data into the hands of North Carolina growers and citizens. Never before has such timely information been so easily accessible by those we serve.
Looking ahead, a secondary impact of the project will be enabling future, non-website delivered data services. Some users may only want the data if a particular threshold is met affecting their operations. A future project could be launched to push specific crop alerts to users who opt for them. Additionally, a dial-in speech-recognition telephone system could be developed to deliver SCO data services to anyone with a telephone.
This project is funded, in part, by the NC State University Office of Extension, Engagement and Economic Development and the North Carolina Agricultural Research Service.
Supplementary URL: http://www.nc-climate.ncsu.edu
Session 6B, Internet Applications and Cyberinfrastructure
Tuesday, 13 January 2009, 1:30 PM-2:45 PM, Room 122BC
Previous paper
Recorded presentation