Search This Blog

Friday, July 6, 2018

Looking for Lightning Data?


As many of you know, the prior BLM lightning viewer (located at https://wfmi.nifc.gov) ceased operation last year.  As a result, two temporary lightning viewer capabilities were created (EGP and a simple BLM).  For those that are not aware of these current/temporary solutions, here are links to those sources:


Temporary BLM viewer:
Also available for those with a WFMI logon (https://wfmi.nifc.gov)

Temporary EGP viewer:
Login; Situation Analysis (SA); OPS; Li; Layers/BLM Lightning

Near term:
EGP has executed a task order to create a new viewer on EGP that is intended to address issues like performance, zooming, caching preferences; timeouts; etc.  This viewer is estimated to be be available in August 2018.  Functions described below.

We ask that users continue their patience with using the temporary solutions until the near-term solution comes online.

Thanks,  John

John Noneman
Project Manager
Bureau of Land Management
National Interagency Fire Center (OC-390)
208-387-5496

***************************************************************************
Instructions:


EGP (Enterprise Geospatial Portal) currently displays Bureau of Land Management (BLM) lightning data in both 2D (Situation Analyst) and 3D (FireGlobe) displays.  Both 2D and 3D displays include a lightning viewstate with layers for BLM lightning data in fixed time ranges: last hour, last 6 hours, last 12 hours, 24 hour, and 2-7 days. The requirements for the viewer are detailed in a document that has been circulating via email, but here is some basic info:

The EGP Lightning Viewer will be an ASP.Net web application hosted by NESS at the EROS data center as a sub-site on the EGP domain similar to FireGlobe and Flight. 
The EGP CCB will serve as the CCB for the Lightning Viewer.

Authentication and authorization will be handled by the EGP LDAP.  All users authorized to view Situation Analyst or FireGlobe will have access to the lightning viewer.


The 2D map display will be the default view.  The default view will be the continental United States (CONUS).  The default display of lightning data will be the previous 12 hours.
The 2D display will have the following Graphical User Interface (GUI) elements/controls:
·         Switch to 3D View
·         Zoom in/out
·         Text display of count of lightning strikes in current view
·         Selection of time period to display
·         Bookmarks for pre-defined geographical areas
·         Layer control to show/hide optional layers
·         Polygon draw toggle.
In addition to the GUI controls, the GUI will support mouse click-drag for panning the map and left-click for opening meta data popups.
The GUI elements are illustrated in Figure 2‑1.  This mockup represents one possible implementation of the required functionality.  The specific interface elements may change during design/development based on requirements and user feedback.  The GUI and underlying code will be based on FireGlobe which will result in the features described below having a similar look and feel to the existing FireGlobe application including its use of headers, footers, fly-out panels and popups.



Lightning strikes will be displayed as individual +/- markers based on the polarity of the strike.  Individual strikes will be shown at all zoom levels.  Testing will determine if there is a number of lightning strikes above which the display does not function properly.  If such a threshold exists, the view will be configured to display an alert with a message containing the number of strikes and a request to the user to change the view, e.g., “The current view contains 110,000 strikes which exceeds the limit of the viewer.  Please zoom in to display fewer strikes for the area of interest.”
Clicking on a lightning strike indicator will open a popup with detail information as shown in the center of the map.
The four buttons at the top left of the map pane will provide access to the layers menu, the bookmarks menu, toggle polygon drawing on/off, and export data.  Each icon will display an activated state.
The table at the top right will provide a count of lightning strikes in the current map view.
The controls at the bottom right will provide zoom, center on user’s location, and toggle to 3D view.
The dropdown at the bottom left will provide time span options including the previous 1, 6, 12, 24, 48 and 72 hour periods and a “Range” option that will display inputs for start and stop time (i.e., date-time input with minute precision).
The polygon drawing feature will allow the user to draw a polygon around a map area.  When the polygon is closed and selected, the lightning strike counts will be updated to display the counts within the enclosed area.  The text in the count table will use a font color corresponding to the polygon line color to provide a visual cue that the counts represent only the polygon area.  When the polygon is de-selected, the counts will revert to showing values for the entire map view.  Polygons will use a set of pre-configured contrasting colors (not user configurable).
Data export will create a file download of CSV data containing time, latitude, longitude, and polarity of strikes in the current view.  Header information will be included with the start and stop time of the time interval selected and the name of the current bookmark if active.



The 3D map display will display the lightning data on a globe view.  The 3D GUI will contain the same controls as the 2D viewer.  The view toggle button at the bottom right will display “2D” instead of “3D” in the 3D view.


EGP currently processes a data feed from BLM containing lightning strike data.  However, because of the volume of lightning data and the limited requirements for lightning strike display in SA and FireGlobe, only seven days of data is retained on a rolling basis.  The current EGP database optimizes query and display by preselecting strike data in the last 24 hours and in the previous two to seven days into separate tables.
The new EGP lightning strike viewer will require access to archived data.  As part of the design process, the database table design will be prototyped to retain efficient processing of data with a larger data set.  In addition, the volume of necessary storage will be estimated from historical data to provide estimates of storage requirements for the future. 

The existing EGP lightning feed processor and database stored procedures will be reused where possible for the new viewer, but the existing EGP process and data will not be changed.  The new EGP Lightning Viewer will use a new ETL process and database tables to make the lightning viewer independent from the current SA and FireGlobe versions.  This will allow SA and FireGlobe to retain their current lightning views or turn them off in the future as appropriate for those applications.

A new API will be designed and implemented as part of the lightning viewer.  The API will allow the client-side viewer to query lightning strike data based on the parameters allowed by the GUI including start and stop time and geographic limits of the current view.

The sections above detail features intended for an initial version-one deliverable.  On-going work may include the following features.  These features will be discussed and prioritized after the initial development, testing and deployment is complete.  The features below are not included in the scope described in Section 4.0.

The initial development will include CSV data format export.  Users may desire the export of spatial data files (i.e., geo-database or GeoPDF).  After the initial API development, the possibility of providing these exports for user-selected time ranges and geographic areas will be investigated.

Initial discussions of functionality included the possibility of having start/stop times set using a graphical slider control.  There are several challenges with using a slider control for start/stop time including:
·         Sufficient time granularity over a large archive data set.  Users will likely require one-hour or less granularity when exploring recent data, but will want to access archive data as well.  If a screen-width slider represents just two years of time span, an hour would be approximately 1/10th of a pixel on typical monitors, making it impossible for the user to perform accurate selections.  A slider control may be most useful for viewing a fixed time range in an animated fashion (e.g., display always shows 6 hours and the slider changes the base time) allowing the user to quickly explore the archive for large-scale lightning events.  However, this use would be secondary to being able to accurately set the start/stop times of interest.  After the initial viewer has been deployed, discussion and prototyping of slider controls can be done based on lessons learned from users and performance of the initial version.
·         The data volume of the archive data may preclude the use of slider controls and fast-changing animations over long time periods.  Typical time periods may contain only tens of thousands of lightning strikes per day, but events with millions of strikes per day must be accommodated.  The initial version of the lightning viewer will allow the developers to accurately assess the archive data size and the data loading requirements of a viewer providing animation of long time periods.
·         Display limitations of a browser-based viewer may limit the number of data points that can be displayed with reasonable performance.  The implementation of a viewer threshold is discussed above in Section 2.2.  The effect of these limitations on the implementation of a slider will be assessed.

Users have requested that the lightning viewer display newly-added strikes in an animated fashion in near-real-time as the data points are loaded from the feed.  Once the basic feed processing and display functionality is in place, the API and rendering requirements for continually updating the display with newly-acquired data will be assessed and tested.

The current lightning viewer in SA and FireGlobe uses lightning strike density representations at zoom levels where the number and resolution of individual strikes is not practical.  If using a heat-map style representation of strike density solves some of the performance issues mentioned in Section 4.2, they will be discussed with users and prototyped as part of the on-going development.



No comments:

Post a Comment