Search Results

You are looking at 1 - 10 of 16 items for :

  • Deep learning x
  • Meteorological Monographs x
  • All content x
Clear All
Sue Ellen Haupt, Branko Kosović, Scott W. McIntosh, Fei Chen, Kathleen Miller, Marshall Shepherd, Marcus Williams, and Sheldon Drobot

machine learning, informatics, pattern recognition, knowledge-based systems, and more ( Smith et al. 2006 ). These nomenclatures attempted to distinguish the more nascent methods that are based more on data from the earlier, primarily heuristic approaches. Industry, however, continued the work and with IBM’s success with Deep Blue beating chess champion Gary Kasparov in 1997, interest in AI resumed ( Smith et al. 2006 ) and U.S. funding agencies began to regain interest in the field. During the boom

Full access
Carl Wunsch and Raffaele Ferrari

historian’s discussion of marine sciences before 1900. Mills (2009) brings the story of general circulation oceanography to about 1960. In the middle of the nineteenth century, the most basic problem facing anyone making measurements of the ocean was navigation: Where was the measurement obtained? A second serious issue lay with determining how deep the ocean was and how it varied with position. Navigation was almost wholly based upon celestial methods and the ability to make observations of sun, moon

Full access
Maria Carmen Lemos, Hallie Eakin, Lisa Dilling, and Jessica Worl

availability of alternative courses of action ( Lemos et al. 2012 ). Furthermore, the very notion of “what use is” in a weather and climate context has been more deeply interrogated and shown to have a variety of dimensions, from use from an instrumental perspective, in which information directly impacts a decision, to use for “enlightenment” or confirmational purposes, in which, although the decision itself may not have changed, information may have played a role in informing or confirming an existing

Full access
Sue Ellen Haupt, Steven Hanna, Mark Askelson, Marshall Shepherd, Mariana A. Fragomeni, Neil Debbage, and Bradford Johnson

patterns and used to train prediction algorithms. One well-used method to predict electrical load is to search for similar days in the past (analogs) and to summarize the corresponding observed daily load curves for use as a prediction. Other methods include statistical or artificial intelligence learning methods, such as regression models, time series methods ( Almeshaiei and Soltan 2011 ), artificial neural networks ( Park et al. 1991 ; Lee et al. 1992 ), and support vector machines ( Hong 2009

Full access
C. N. Long, J. H. Mather, and T. P. Ackerman

change, and the effects of climate change relevant to the region; and 3) foster career goals in science for students in the region. It organized workshops for teachers in both countries, working with several regional partners in the South Pacific island nations and Australia. It also worked to develop curriculum units and a variety of teaching tools, including a traveling kiosk that could be used for interactive learning and viewing the data. In the mid-2000s, in response to a general lack of funding

Full access
M. A. Miller, K. Nitschke, T. P. Ackerman, W. R. Ferrell, N. Hickmon, and M. Ivey

employed in the AMFs, while in other cases not. And the increasing breadth and sophistication of the AMF’s measurement suite has challenged the knowledge base of even the savviest scientist. So the position of site scientist has changed with the AMF1 and in the latter years the focus has been upon learning new instrumentation, performing ASR cloud and radiation science when the experiment focus did not address the basic programmatic needs of each deployment, and performing high-level quality control. 6

Full access
David A. Randall, Cecilia M. Bitz, Gokhan Danabasoglu, A. Scott Denning, Peter R. Gent, Andrew Gettelman, Stephen M. Griffies, Peter Lynch, Hugh Morrison, Robert Pincus, and John Thuburn

exchange energy with the rest of the universe, and because motions of the atmosphere are fundamentally driven by spatial gradients in the electromagnetic radiation emitted by Earth, its atmosphere, and the sun. The same gradients also play a key role in determining the thermal structure of the atmosphere. The deep convective clouds of the tropics arise from a rough balance between destabilization by radiative cooling and the response of deep convection, for example, while the planetary-scale Hadley

Full access
Thomas P. Ackerman, Ted S. Cress, Wanda R. Ferrell, James H. Mather, and David D. Turner

this period, the Science Team was actively engaged in learning how to use effectively the available and expected data. Initially, the ARM Science Team was organized around two broad research themes. One was the instantaneous radiative flux (IRF) concept and the other was the single-column model (SCM) approach. The IRF approach was derived from the SPECTRE experience ( Ellingson et al. 2016 , chapter 1) and focused on measuring all the radiatively active components in an atmospheric vertical column

Full access
A. Korolev, G. McFarquhar, P. R. Field, C. Franklin, P. Lawson, Z. Wang, E. Williams, S. J. Abel, D. Axisa, S. Borrmann, J. Crosier, J. Fugal, M. Krämer, U. Lohmann, O. Schlenczek, M. Schnaiter, and M. Wendisch

sample volume. By numerically reconstructing each hologram ( Fugal et al. 2009 ), one obtains the three-dimensional positioning and imaging of each individual particle. An example of a particle spatial distribution, along with the size distribution and image gallery reconstructed from a single hologram, is presented in Fig. 5-11a . The particle images are classified via supervised machine learning as described in Schlenczek et al. (2017) . Figure 5-11a provides insight on how ice particles and

Full access
Stanley G. Benjamin, John M. Brown, Gilbert Brunet, Peter Lynch, Kazuo Saito, and Thomas W. Schlatter

intervention in critical weather and environmental situations and improved techniques for forecasting the impacts of weather-related hazards. The information from these forecasts will feed applications tailored to the needs of end users. In light of the complexity and the huge size of the dataset generated by next-generation NEWP systems, it is evident that artificial intelligence (e.g., deep learning) and machine-learning techniques will be used routinely to integrate forecasts into the decision

Full access