A Blend of Science, Art, and Intuition: An Overview of the Mauna Kea Weather Center

Ryan Lyman, Tiziana Cherubini, and Steven Businger

Mauna Kea Weather Center, Institute for Astronomy, Hawaii, U.S.A.

The primary mission of the Mauna Kea Weather Center (MKWC) has always been to provide custom forecasts to optimize astronomical operations on the summit of Mauna Kea. (The MKWC is comprised of Dr. Tiziana Cherubini - Research Meteorologist in charge of numerical modeling, Ryan Lyman - Forecast Meteorologist who provides most of the daily custom forecasts and oversees product delivery, and Dr. Steven Businger - PI who provides program direction.) Custom forecasts are created by integrating current and past observations around the Big Island, with output from global and mesoscale numerical models. A custom mesoscale modeling system (based on the Penn State/NCAR mesoscale model MM5) is run twice daily and provides high-resolution 60 hour forecasts for the Mauna Kea summit and Hawaii (Figure 1).

Observational data relied on by the forecasters include remotely sensed data (those collected by satellites, radars, and lightning detectors) and in situ data collected by standard surface instruments on land and on ships, as well as radiosondes released twice daily from Hilo and Lihue. An objective analysis program is used to combine or assimilate the in situ and remotely sensed data into a form that can be ingested by the MM5 model. Forecasters rely on the observational data streams for making short-term forecasts (out to ~12 hours), called nowcasting, and for validation of the model output. Although nowcasting may seem straightforward, integration of in situ data and interpretation of satellite and radar imagery takes years of experience to master. Forecasts out to 60 hours require an intimate knowledge of the strengths and limitations of complex numerical models, as well as a comprehensive understanding of the dynamics involved in the interplay between atmospheric circulation systems, the ocean, and the complex island terrain. In the end, custom forecasts are derived not only through pure science, but also through a blend of art, experience, and intuition!

Figure 1: Graphical Forecast Trend for the Friday August 11, 2006, morning forecast.

In order to reliably deliver these forecasts, a suite of custom and operational meteorological data and products are constantly retrieved, generated, posted, and archived on the MKWC web server. While most of these programs are automated, the MKWC staff vigilantly checks for properly running data feeds, numerical models, and servers, both within the MKWC infrastructure and at Subaru (which hosts the mesoscale model runs). Moreover, we are regularly developing or restructuring our products to keep up with the changing needs of the observatories and new developments in both observations and numerical models. A significant effort is devoted to ensuring a consistent, stable, and advanced operation that the observatories and the support staff can count on.

Figure 2: Projected MM5 seeing distribution and winds for the surface, focused on the summit area for Saturday August 12, 2006, 0200 HST.

Extensive research is being done to develop an accurate and reliable seeing and refractive index (Cn2) profile forecast. An optical turbulence algorithm has been implemented that makes use of the information on turbulence kinetic energy provided by the mesoscale model along with the model predicted winds, temperature, and moisture gradients. Cn2 profiles and 2-D maps of seeing are posted on the MKWC web pages (see e.g. Figure 2), and are used as guidance by MKWC forecasters when issuing seeing forecasts. Observed data obtained by the different observatories and during site-monitoring field experiments are used in an ongoing effort to validate and refine the seeing algorithm.

The MKWC is very much at the forefront of research and development of operational seeing forecasts. More details are given in the publications listed below.

References:
Cherubini, T., S. Businger, C. Velden, and R. Ogasawara, 2006a: Assimilation of satellite derived winds in mesoscale forecasts over Hawaii. Mon. Wea. Rev., 134, 2009-2020.
Cherubini, T., S. Businger, R. Okasawara, and R. Lyman, 2006b: Modeling turbulence and seeing over Mauna Kea. Submitted to J. Appl. Meteo., In review.