"Weather-Impact Decision Aids Software to Help Plan"
Weather-Impact Decision Aids: Software to Help Plan Optimal Sensor and System Performance Dr. Richard C. Shirkey Melanie Gouveia Army Research Laboratory Northrop Grumman Weather can play a decisive role in military battles, in their planning, and in their execution. Weather-impact decision aids give the commander an edge by allowing both a determination of the optimum selection of weapon systems and a comparison with threat systems under current or forecast weather. This article describes two weather-tactical decision aids: the Integrated Weather Effects Decision Aid and the Target Acquisition Weapons Software. W eather is ubiquitous; planning for it is an everyday occurrence, yet it still manages to foul up our plans. Recent mili- and threat) has its list of relevant rules, which include red-amber-green (unfavor- able-marginal-favorable) critical value thresh- map overlays (see Figure 3, page 18) for the region of interest. Environmental data for the region of interest is supplied primarily tary examples abound, such as dust clouds olds for one or a combination of the envi- via the Army’s Battlescale Forecast Model that grounded sorties in Operation Allied ronmental parameters that affect the sys- , developed for short-range forecasting. Force in Kosovo. To effectively execute tem. Results are displayed via a matrix of The environmental impact rules and critical missions, the military commander must be impacts vs. time (see Figures 1 and 2) and values for the various systems have been aware of the weather and its impact on Figure 1: IWEDA Weather Effect Matrices his/her equipment, personnel, and opera- tions. There are a number of weather- impact decision aids (WIDAs) that deter- mine weather effects on mission-selected equipment and operations. Generally, these WIDAs may be broken into two subsets: rule-based and physics-based. Rule-based WIDAs, such as the Army’s Integrated Weather Effects Decision Aid (IWEDA) , are constructed using observed weather impacts that have been collected from field manuals, training cen- ters and schools, and subject matter experts. IWEDA provides information (in the form of stoplight charts) concerning which weapon systems will work best under fore- cast weather conditions; no information is provided concerning target acquisition range. Physics-based tactical decision aids (TDAs), such as the Tri-Service Target Acquisition Weapons Software (TAWS) , employ physics calculations that have their Figure 2: IWEDA Full Impacts basis in theory and/or measurements. TAWS determines the probability of detect- ing a given target at a given range under existing or predicted weather conditions. Thus, physics-based systems produce results in terms of a performance metric that take on a continuum of values rather than the simpler stoplight results from the rule-based systems. The IWEDA IWEDA, a UNIX-based program written in Java, is a collection of rules with associated critical values for aiding the commander in selecting an appropriate platform, system, or sensor under given or forecast weather conditions. It provides qualitative weather impacts for platforms, weapon systems, and operations, including soldier performance. Each system (Army, Air Force, Navy, December 2002 www.stsc.hill.af.mil 17 Software Engineering Technology causes a significant (moderate or severe) systems, their subsystems, and components. impact on a military operation, system, sub- Results are presented as a function of time system, or personnel. and location. In general, the rules are determined by To construct the example mission, the operational usage (as embodied in the field A-10, AH-64, personnel, SA-14, and SA-16 manuals, etc.), whereas the critical values are systems were selected from IWEDA’s friend- determined by doctrine, safety, or engineer- ly and threat graphical user interfaces (GUI). ing factors (people, modeling, or testing). Once these systems have been selected, Currently IWEDA stores information on IWEDA determines the weather impacts 102 systems, 86 of which are friendly, 16 of on the mission; results are presented to the which are threat-rated. user in the form of a WEM, as shown in Figure 1. IWEDA Operational Usage Initially, the lower half of the WEM is IWEDA is arranged in a fashion that pres- blank with the upper half showing the ents systems, subsystems and components weather impacts as a function of system(s) in a hierarchal fashion. A group of systems and time. By performing a right click on any Figure 3: IWEDA Map Overlay for AH-64 is called a mission; a system often contains of the colored cells, such as the AH-64 for TADS TV/DVO one or more subsystems; the subsystems 22/12 (day/time), condensed impacts are often have one or more components. The shown in a scrollable window in the lower validated through the Training and half of the WEM (impacts for the config- user has the option to define which systems Doctrine Command’s organizations, field belong to a mission and to delete optional ured AH-64 system have been reproduced manuals and the National Ground and subsystems and components from a system in Table 1). The WEM shows impacts on Intelligence Center. thereby allowing a determination of weath- the AH-64 system as a function of time and IWEDA is currently being fielded as er impacts from operations or missions at general environmental conditions, but we part of the Army’s Command, Control, the highest level down to systems, subsys- do not know the full (detailed) impact or Communications, Computers and Intelli- tems, and components at the lowest level. where the impact is occurring within the gence (C4I) tactical weather system, the For missions, systems, subsystems, and forecast area. Integrated Meteorological System. As a C4I components, the impacts over the forecast To determine the full impact statement tool, IWEDA does not dictate a course of period are shown on weather effects matri- and the location, a left mouse click is per- action, but only informs the commander ces (WEMs, see Figure 1). The WEM is formed on the AH-64 cell for the selected that there will be weather impacts on the color-coded; for use with non-color print- day of the month and time, i.e., 22/12. This force (friendly or threat). ers, cells are annotated with R (red), A brings up the next screen (see Figure 2) that IWEDA rules, which interact with the (amber), or G (green). Red areas indicate presents all of the selected AH-64 subsys- weather database to determine impacts on that operations are severely impacted: There tems and components and their color- the selected system(s), are determined from is either a total or severe degradation or the coded impacts. system concepts and are embodied in a operational limits or safety criteria have As in the WEM GUI, initially only the computer database that has been tied to crit- been exceeded. Amber indicates that opera- top half of Figure 2 is presented to the user. ical values. The critical values are defined, in tions are marginal and the operational capa- To obtain further information, the user a meteorological sense, as those values of bility is degraded, or there is a marginal clicks on one of the colored blocks; in the weather factors that can significantly reduce degradation. Green indicates that there are example presented, the TV/direct view the effectiveness of, or prevent execution no operational restrictions. sight component of the Target Acquisition of, tactical operations and/or weapon sys- Based on requirements, users may query Designation Sight (TADS) has been inter- tems. and view various levels of information: text rogated. This results in a color-coded map An example of such a rule would be impact statements or spatial distributions of overlay (Figure 3) showing where the “usage of TOW2 is not recommended for impacts on a map overlay. TV/D is affected by the weather. The full visibilities less than three kilometers.” In impact statement, along with its source, can this example rule, a visibility of three kilo- IWEDA Example now be obtained by moving the cursor meters (the critical value) has been coupled In the following example, a user-defined (shown as a white circle) and clicking upon with a system (TOW2) resulting in a rule. mission is created by selecting three friend- a white area on the map (upper left of cen- We can further define this critical value, or ly and two threat systems. Once the mission ter). range of values, as the point where the has been configured, the database is queried The associated full impact statement occurrence of a meteorological element to determine the weather impacts on the then appears in the lower half of Figure 2, which in this case is “Any occurrence of fog Table 1: Impacts for the AH-64 System for 22/12 or visibility <1.9 mi (3100m) significantly System AH-64 has marginal impact: High Pressure Altitude reduces the target and background contrast Subsystem 30 MM MACHINE GUN has marginal impact: Low Visibility making target acquisition difficult.” Component THERMAL SIGHT has marginal impact: Reduced Visibility Contrast this with the condensed impact Component TV/DIRECT VIEW SIGHT has marginal impact: Reduced Visibility statement of “Fog and Low Visibility” Component TV/DIRECT VIEW SIGHT has unfavorable impact: Fog and Low Visibility shown in the WEM. Component Laser R/D has marginal impact: Reduced Visibility In summary, the colored cells in the Component Laser R/D has unfavorable impact: Low Visibility WEM display the worst-case condition for the Subsystem HELLFIRE has marginal impact: Icing Aloft selected mission, during the selected time, for Subsystem GENERATOR has marginal impact: High Altitude the entire forecast region. If the user wishes to Component NIGHT VISION GOGGLES has unfavorable impact: Reduced Illumination know why a particular cell is red or amber, 18 CROSSTALK The Journal of Defense Software Engineering December 2002 Weather-Impact Decision Aids: Software to Help Plan Optimal Sensor and System Performance further information is available in impact position), contrast at visible wavelengths, by the Multi-Service Electro-optic Signature statements, which explain why a particular C(0), is the difference between the target, model (MuSES) , which calculates the cell exists. Detailed analysis for the impact- It , and background, Ib, radiances, divided equilibrium background and target temper- ed system or sensor can be obtained from by the background radiance, atures using antecedent illumination and the color-coded map. weather data. C(0) = [It(0) — Ib(0)]/Ib(0). (1) MuSES has two primary components: a The TAWS thermal analyzer module and a signature TAWS , a GUI-based program running We may express the apparent contrast at model. Thermal analysis is the computation under the Windows operating system, is a range r as of physical temperature and heat rates that Tri-Service program that includes Air C(0) are obtained through energy balance on a Force, Army, and Navy sensors and targets. C(r) = ___________________ , node or isothermal element using a finite- TAWS supports systems in three regions of 1+[ Ip(r)/Ib(0)][ 1/T(r)] (2) difference numerical solution of the differ- the spectrum: visible (0.4 - 0.9 microns), ential equations. The main output of a ther- laser (1.06 microns), and infrared (IR) (3-5 where T(r) is the atmospheric transmission, mal model is physical temperatures and net microns; 8-12 microns). It accepts current and Ip is radiation scattered from atmos- heat rates that compare to empirical meas- or forecast weather data to determine target pheric aerosols and gases into the line-of- urements of contact sensors. detection range for selected sensors and tar- sight. Ip is called the path radiance and may The signature analysis is the computa- gets. The commander uses this information be thought of as atmospheric noise scat- tion of apparent temperature or radiance, for mission-planning purposes or to ascer- tered into the sensor’s field of view; it is not which is composed of an emitted compo- tain which sensors can see the furthest dependent upon the target. nent that is a function of physical tempera- under the given weather conditions. In TAWS at visible wavelengths, the tar- ture and emissivity and a reflected compo- TAWS performs both illumination and get and background radiances are deter- nent that is a function of irradiance from its performance prediction calculations (PPC). mined using Hering and Johnson’s Fast surroundings and its reflectivity. In other The PPC can be done for single or multiple Atmospheric SCATtering model (FASCAT) words, the signature is what a sensor views locations during a mission. The illumination , which calculates upwelling and down- and measures the radiance of a target, analysis involves the computation of solar welling radiance terms at specified heights which is only partially dependent on its and lunar ephemeris information for a spec- in the atmosphere. physical temperature. Thus, the signature ified location. A mission planner, for exam- For designated sensor and target alti- model provides a link between the output ple, might be interested in an illumination tudes, the apparent contrast is calculated for of the thermal model and the desired out- analysis to determine the time of sunset for slant paths, which may include an optional put in signature analyses. a particular mission date and location. For a cloud layer. Objects in sunlight or shadow The basic heat source components con- single location, the PPC could be used to may be viewed against sky, cloud, or terrain sidered by MuSES include longwave radia- predict detection range for a particularly backgrounds. The path radiance Ip, and the tion, solar absorption, engine heating, important target as a function of time, background radiance Ib, are determined by engine compartment air, exhaust gas, track while a PPC for multiple locations along a a multiple scattering calculation using the and wheel heating, and convection. Inter- mission route would be useful to a mission delta-Eddington approximation  in con- reflections between diffuse surfaces are also planner predicting detection ranges for a junction with the atmospheric model. The taken into consideration. These various series of key locations as a function of time. contrast is subsequently determined using temperatures and effects are used to calcu- To determine the acquisition range to a equation (2). late ∆T. given target a number of quantities need to For visible/near-IR scenarios, the target be known: the target-to-background con- may be on the ground or elevated. An ele- Laser Contrast Model: The laser model trast, the atmospheric conditions, solar or vated target may be viewed with an upward does not compute contrast. lunar luminance, and sensor characteristics, or downward line-of-sight (LOS). Sky and all of which vary with spectral region. We cloud backgrounds are supported for the Atmospheric Information discuss each of these in the following sec- upward LOS; distant earth and low-lying To determine the loss of energy as radiation tions and provide an illustrative example at cloud backgrounds are supported for the passes through the atmosphere requires the end. downward LOS. knowledge of the atmospheric constituents (gases and aerosols) and its state (pressure, Target-to-Background Contrast Thermal Contrast Model: The inherent temperature, relative humidity, etc.). This Contrast is defined as the ability of an contrast at thermal wavelengths is defined loss of energy is expressed in the form of observer to distinguish an object from its as the target temperature minus the back- atmospheric transmission, which ranges in background; it degrades as the atmospheric ground temperature, value from zero to one and is highly path length increases. At visible wave- dependent upon the aerosol type present. lengths, where radiation scattering from C(0) = [It(0) — Ib(0)] = ∆T, (3) This loss of energy can be represented by atmospheric particulates is important, the Beer’s law for atmospheric transmission, mathematical formulation of the contrast is where ∆T is the temperature difference different than in the infrared (IR), where between the target and background. Note T(r) = e- (ka + kp + km) r, (5) emission is the dominant process. Since that as the temperature increases, so will the TAWS computes contrast in both of these inherent radiance, I(0). Thus, the contrast in where ka, kp, and km are the aerosol, pre- spectral regions, we present the following the IR is, cipitation, and molecular extinction coeffi- formulations. cients, respectively. The molecular extinc- C(r) = C(0) T(r) = ∆T T(r). (4) tion coefficients are determined in TAWS Visual Contrast Model: The inherent, or by using a scaled down version of the low zero range (usually defined as the target’s In TAWS, C(0) is determined indirectly transmission atmospheric propagation code December 2002 www.stsc.hill.af.mil 19 Software Engineering Technology Aerosol Properties ible and infrared spectral bands. Ranges and Rural Boundary layer background aerosol found in continental air masses. probabilities predicted by the model repre- Urban Rural aerosol plus an added component representing soot-like sent the expected performance of an aerosols that include particles produced in urban and industrial ensemble of trained military observers with complexes. respect to an average target having a speci- Maritime Characterizes aerosols that include sea-salt particles; the target area is more than a few kilometers inland. fied signature and size. TAWS currently Tropospheric Characterizes aerosols found in very clean air masses and in the free only supports detection ranges; other acqui- atmosphere above the boundary layer. sition ranges are scheduled to be added in Desert Characterizes aerosols found in the boundary layer of desert, arid, or the near term. semi-arid climatic regions. TAWS supports two different classes of Navy Maritime Describes aerosols found in the boundary layer of oceanic systems that employ laser designators oper- environments; includes wind speed dependence. ating at 1.06 microns: laser ranging and laser Advective Fog Characterizes wet aerosols found in dense fogs, where visibility is less than 1 km. lock-on systems. Each of these has designa- Radiative Fog Describes aerosol properties in less dense fogs, where visibility is 1 km tor and receiver components. The airborne or greater. laser ranging systems measure the distance Camouflage Smokes Characterizes white phosphorus, fog oil, and hexachloroethane from the ranger system to the target by smoke. measuring the travel time of the laser pulse Battlefield Induced A persistent pall of smoke and dust that sometimes covers areas from the designator to the target and from Contaminants (BIC) where intense combat has occurred. the target to the receiver. The designator Table 2: TAWS Aerosol Models and receiver are physically collocated in the LOWTRAN . The aerosol extinction itation type/rate; surface aerosol type; bat- same hardware package for all ranging sys- coefficients [8, 9] are read from pre-calcu- tlefield induced contaminants; high-, mid-, tems. For the laser lock-on weapons, the lated internal tables. and low-level clouds*; and the boundary designator illuminates the target and the TAWS contains 10 aerosol and two pre- layer height. receiver receives the reflected beam. TAWS cipitation models that are used in various predicts the maximum effective range for combinations by the IR, television/night Solar/Lunar Illumination either the designator or lock-on receiver. vision goggles, and laser models to deter- Illumination analysis in TAWS involves the mine the appropriate aerosol and/or pre- computation of solar and lunar ephemeris Example cipitation extinction coefficients. The data for a specified location and a series of We present here a winter scenario using a T- aerosols describe the primary particulates of dates or times. Solar/lunar ephemeris input 80 Soviet main battle tank in exercised and the air mass close to the surface at the loca- information is derived from user-input time off modes, against a snow background at tion of interest. The naturally occurring of day/time of year and latitude/longitude, IR wavelengths. The sensor and tank were aerosols include rural, urban, maritime, tro- in conjunction with the Solar-Lunar aligned such that the sensor always had a pospheric, desert, advective fog, radiative Almanac Code . frontal view of the tank; the sensor height was 10 feet. The date and location were fog, and Navy maritime. There are three The solar/lunar ephemeris information fixed at 21 December, latitude 37° 32’ N, types of camouflage smokes: white phos- is also computed and used for target acqui- longitude 127° 00’ E (Seoul, S. Korea), phorus, fog oil, and hexachloroethane. A sition analysis. In this case, in conjunction respectively. The weather conditions were 10th aerosol, in the form of battlefield with variable cloud cover, the solar/lunar overcast and snowing with visibilities of induced contaminants, is available for situa- position is used to calculate target/back- three miles (light snow) and one mile (heavy tions where there is a persistent pall of ground heating for the IR model and inher- snow) with a light breeze (~3m/s) from the smoke and dust raised by combat. ent target/background radiance for the vis- west. The relative humidity and tempera- Properties of the aerosol models are pre- ible model. The laser model does not use ture, taken from a climatological database sented in Table 2. TAWS also contains rain ephemeris information. , as a function of local time are present- and snow precipitation models. ed in Table 3. TAWS allows a wide range of meteoro- Sensor Information The results of the model run are shown logical conditions, all of which may be Sensors are user-selected once the spectral in Figure 4. The two vertical lines, deter- selected by the user and some of which may region has been chosen. The relevant sensor mined using the illumination analysis capa- be automatically input via the Air Force curve is automatically retrieved from the bility of TAWS, indicate the sunrise and Weather Agency (AFWA) or the Navy sensor database. sunset times. As expected, the detection Tactical Environmental Data Server Within TAWS, target detection range range is considerably larger when the visi- (TEDS). These meteorological parameters for Silicon TeleVision (TV), night vision bility is higher; for given weather conditions include the following (those values noted goggles (NVG), and IR sensors is deter- the exercised tank is easier to detect relative with an asterisk may be downloaded from mined by using the Acquire sensor per- to the tank in the off state. AFWA or TEDS): atmospheric dewpoint formance model . Acquire predicts tar- Thermal crossover, defined as the time temperature*; sea surface temperature*; get detection and discrimination range per- during the day when the thermal contrast is wind velocity/direction*; visibility*; precip- formance for systems that image in the vis- at a minimum and the polarity of the con- Table 3: Input Relative Humidity (RH) (%) and Temperature (°C) as a Function of Time (HRS) trast reverses, generally occurs at mid- morning and late afternoon. For example, Time 1800 2100 0000 0300 0600 0900 1200 1500 1800 2100 0000 in early morning the background tempera- ture may be greater than the target temper- Temp 1 -3 -3 -5 -5 0 0 1 1 -3 -3 ature. After thermal crossover, the target temperature may be greater than the back- 69 74 74 86 86 80 80 69 69 74 74 RH ground temperature. In the example, ther- 20 CROSSTALK The Journal of Defense Software Engineering December 2002 Weather-Impact Decision Aids: Software to Help Plan Optimal Sensor and System Performance mal crossover occurs at approximately 0900 Software Products for Operational Light snow, 3 mi visibility, Tank on Light snow, 3 mi visibility, Tank off and 1700, accounting for the low detection Weather Support. Proc. of the Heavy snow, 1mi visibility, Tank on Heavy snow, 1 mi visibility, Tank off range at those times. The commander/user Battlespace Atmospheric and Cloud 5 can now optimize assets by choosing a time Impacts on Military Operations Con- when detection range is maximized and by ference. Fort Collins, CO, Apr. 2000. avoiding those times such as when thermal 4. Hering, W.S., and R.W. Johnson. The 4 crossover occurs, when detection ranges are Detection Range (km) FASCAT Model Performance Under at a minimum. Fractional Cloud Conditions and 3 Using this information in conjunction Sunset Related Studies No. 84-0168. University with weather forecast information (as of California, Scripps Institution of Sunrise opposed to static information used in this Oceanography. San Diego, CA, 1984. 2 example) provides additional relevant infor- 5. Joseph, J.H., W.J. Wiscombe, and J.A. mation. For example, let us examine the Weinman. “The Delta-Eddington 1 “tank on” curves in Figure 4. If the weath- Approximation for Radiative Flux er conditions were predicted to change Transfer.” JAS Vol. 33: 2452. from heavy to light snow at 1200 local, the 6. Johnson, K., et.al. MuSES: A New Heat 0 detection range would increase from and Signature Management Design Tool 0 4 8 12 Local Time (hrs) 16 20 24 approximately one and one-half kilometers for Virtual Prototyping. Proc. Ninth to approximately four and one-half kilome- Figure 4: Detection Range vs. Time Annual Ground Target Modeling & ters, providing the commander with an Validation Conference. Houghton, MI, DAT.” ASL Technical Report 0160-3. opportunity for increased detection. Such Aug. 1998. July 1986. scenarios may also be used for 7. Kneizys, F.X., et. al. “Atmospheric 10. Bangert, J.A. Solar-Lunar Almanac friendly/threat comparisons to determine Transmittance/Radiance: Computer Code (SLAC) Software User’s Guide the delta in range due to differing systems. Code LOWTRAN 6.” Air Force Version 1.1. U.S. Naval Observatory, Geophysics Laboratory Technical Astronomical Applications Depart- Conclusions Report 83-0187. Hanscom AFB, MA, IWEDA provides the commander with an ment, 1998. easy-to-use and interpret tactical application 1983. 11. U.S. Army. Acquire Range Performance that allows for near real-time evaluation of 8. Shettle, E.P., and R.W. Fenn. “Models Model for Target Acquisition Systems. sensor employment options. Automating for the Aerosols of the Lower Version 1 User’s Guide. U.S. Army the environmental parameter retrieval by Atmosphere and the Effects of Humidity Variations on Their Optical CECOM Night Vision and Electronic using a prognostic data set further enhances Sensors Directorate Report, Ft. Belvoir, the application and allows for realistic plan- Properties.” Air Force Geophysics Laboratory Technical Report 79-0214. VA, 1995. ning based on evolving weather. Hanscom AFB, MA, 1979. 12. Avara, E., and B. Miers. “The TAWS aids the warfighter in determin- ing what sensor/weapon system will work 9. Shirkey, R. C., R. A. Sutherland, and M. Climatology Model CLIMAT.” Army best against a user-selected target under A. Seagraves. “EOSAEL 84: Vol. 3, Research Laboratory Technical Report adverse weather conditions. TAWS accom- Aerosol Phase Function Database PFN- 273-8. Adelphi, MD, June 1998. plishes this by using accepted sensor per- formance and aerosol models coupled with About the Authors proven techniques for determining atmos- Richard C. Shirkey, Melanie Gouveia pheric transmission and contrast. In addi- Ph.D, is a physicist with manages the Weather tion to determination of acquisition ranges, TAWS may be used for mission planning the Army Research Impact Decision Aid and for determination of deltas between Laboratory’s Compu- projects at Northrop friendly and threat systems. tational and Infor- Grumman. She has Taken together, these TDAs provide the mation Sciences Direc- worked in the areas of commander a significant advantage for sys- torate, Battlefield Environment atmospheric modeling and tactical tem selection under adverse weather condi- Division. He has worked in the area of decision aids for the past 12 years. She tions.◆ atmospheric modeling and simulation is currently leading model improve- for the past 23 years and is currently ment, graphical user interface develop- References engaged in atmospheric effects for tar- ment, and environmental data source 1. Sauter, D., and R. C. Shirkey. Target Ac- get acquisition and their impacts on efforts for the Target Acquisition quisition Modeling with Automated wargames. Weapons Software effort. Environmental Data Ingest for Weapon System Evaluation. Proc. of Ground US Army Research Laboratory Northrop Grumman Target Modeling & Validation AMSRL-CI-EE Information Technology Conference. MI, Aug. 1999. White Sands Missile Range, NM 55 Walkers Brook Drive 2. Henmi, T., R. Dumais Jr. “Description 88002-5501 Reading, MA 01867 of the Battlescale Forecast Model.” Phone: (505) 678-5470 Phone: (781) 205-7202 Army Research Laboratory Technical Fax: (505) 678-4449 Fax: (781) 942-2571 Report 1032. White Sands Missile Range, NM, June 1997. E-mail: firstname.lastname@example.org E-mail: mgouveia@northrop 3. Gouveia, M.J., et. al. TAWS and NOWS: grumman.com December 2002 www.stsc.hill.af.mil 21