weather map

Two Fundamental Principles for Developing Algorithms in an Enterprise Ground System

Allan Weiner, Ph.D., Harris Corporation Senior Scientist
Oct 7, 2016

The launch of National Oceanic and Atmospheric Administration’s (NOAA’s) next-generation satellite, GOES-R, reminds us just how far we’ve come in weather sensing and data processing technology over the past decade. GOES-R will offer three times more spectral information, four times the spatial resolution, and more than five times faster coverage than the current GOES system. As a meteorologist, I find that I’m more excited about the benefits we’re going to see than I’ve been about anything in my 38-year career.

The next important step for NOAA will be development of an enterprise ground solution for the organization’s entire weather enterprise—a single ground system that serves not only GOES-R, but also other NOAA missions, including legacy polar, geo, and commercial Systems, as well as unnamed future ones.  Certainly there will always be unique missions with unique products and needs, but overall, a single infrastructure that can eliminate hardware, software, and reduce personnel seems like a no-brainer when it comes to increasing efficiency and saving money.

Unfortunately, evolving an enterprise ground system from disparate systems is not necessarily as easy as it sounds. Each weather mission has its own set of objectives and associated algorithms.  A single ground system will require a common and, more significantly, a new set of algorithms. How can these new algorithms be efficiently moved into a new operational enterprise system—and in a way that enables them to work successfully for all missions? Here are two fundamental principles to guide the process.


Researching and developing algorithms for NOAA’s weather missions and then transferring them to operations has typically involved three entities. A government science team who “owns” and oversees algorithm science, authoring the algorithms. Industry partners who develop and implement the operational algorithm software/system.  Finally, government leadership who serves as the go-between for these two groups, providing process oversight, verifying user requirements. They also ensure that algorithm requirements are consistent with industry requirements, and provide an operational algorithm software/system architecture that is synergistic with the algorithm requirements.

Creating a common set of algorithms for an enterprise ground system would benefit from a fourth entity: an integrated integration and development (I&D) team. The I&D team provides a forum for the government science team and industry partners to come together to communicate and coordinate their activities and technical concerns under the oversight of government management. This enables direct communication between the industry engineer implementing the algorithm and the algorithm science author.


For the I&D team to be successful, a few basic conditions should be met:

  • Team members have a common mission perspective and are committed to achieving program objectives.
  • A solid set of rules clearly defines responsibilities—and especially, who is responsible for algorithm science (research) and who is responsible for implementing the science (operations).  Note that although responsibilities are clearly defined, researchers are still encouraged to help with implementation and implementers are encouraged to suggest alternative science; in this way a true team spirit is built.
  • Communication processes between the groups and rules of engagement for meetings are established and understood by all.
  • The government science team, industry partners, and government management each have a single designated manager with total authority to control all facets of the algorithm work that is within their scope.
  • Algorithm theoretical basis documents to bridge requirements and science are clearly written and understood by all.
  • Individual measurements of success are defined for the government science team, industry partners, and government management.

By organizing the algorithm development team for successful collaboration and then making sure that their products are designed for both research and operations, NOAA will be one step closer to executing a successful transition to a weather enterprise ground system. This methodology was used during the successful creation of the GOES-R ground system—and the I&D team was a major reason these algorithms were implemented on time and within budget!

Read more about Harris’ involvement in the GOES-R mission.