The Effects of Ducting on an X-Band, Phased Array Radar ...



The Effects of Ducting on an X-Band, Phased Array Radar: An AREPS Case Study

LCDR Thomas B. Keefer

OC3150

15SEPT2006

The Effects of Ducting on an X-Band, Phased Array Radar: An AREPS Case Study

I. Introduction and Background

Navy METOC Officers assigned to various deployable assets are routinely called upon to make critical decisions about variations in the capabilities of their unit’s sensors (detection), as well as those of their enemy’s (counter-detection). Assuming an acceptable state of material readiness and state of repair for the organic sensor suite, and having confidence in the capability of intelligence officers to adequately classify the sensors being utilized by enemy forces, the primary source of performance variability exhibited by these systems is caused by environmental effects. Decisions affecting mission timelines, asset allocation and even go/no-go criteria are ostensibly based upon characterization of the environment. Although the pressure of a can-do attitude among senior level mission planners can affect the decision making process, especially during situations when the METOC Officer is suspect of the quality of a model output, this is a separate issue. It is assumed during this study that all decisions are made free from this pressure and based solely on environmental characterizations.

As operational tempo increases in littoral regions of the world, understanding and accurately predicting environmental effects on radar systems has become more critical. Mine warfare (MIW) and Naval Special Warfare (NSW) assets conduct the vast majority of their maritime operations in the littoral and are often required to operate close to or within the range of enemy sensor systems. Operations inside of enemy sensor networks have serious implications for not only the unit conducting the operations, but can potentially jeopardize the element of surprise or endanger the security of larger units. METOC Officers are trained in the use of Tactical Decision Aids (TDAs) such as AREPS, which offer a good first guess about what a sensors performance might be given a set of environmental conditions. These TDAs can be very useful, however heavy reliance in their output, or blind faith in the sensors gathering the input data they are initialized with, can lead to disastrous results. It is imperative that the output of such TDAs be scrutinized for sensitivity to variables that have suspect accuracy and/or precision, whenever possible, to the various inputs they base their decisions on in order that an accurate assessment of the confidence of this output can be made.

Among the myriad of information available in AREPS, the most widely used is the prediction of effective radar ranges. This is particularly useful in detection/counter detection. Over land, AREPS normally performs quite well, but the unique temperature and humidity profiles that exist over the ocean in many part of the world lead to drastic underestimations in radar performance characteristics. Some possible error modes, particularly the errors in radar range prediction associated with the effect of evaporative ducting, can have enormous impact on TDA output validity and can be quite difficult to accurately predict. This is particularly the case when dealing with high frequency fire control and surface search radars such as the X-band radar evaluated in this study. In order to effectively address this particular difficulty, a series of experiments conducted off the coast of Dam Neck, VA and at San Clemente Island, CA (Davidson, et al., 2003, Keuhn, 2003, Moneymaker, 2005,). The experiments included operating an NSW 11m RHIB in a simulated enemy waterway. Two separate radar systems were used; the first was a high frequency X-Band radar, while the other was a lower frequency S band radar. The study detailed here serves as a follow on to the previous two studies, and utilizes the same techniques used by Davidson (2003) in order to examine the performance of an X-Band, phased array radar radar. The study was conducted in July of 2006 in Southern Monterey Bay, CA using the RV Pt Sur as a simulated, medium sized Naval Asset and represents an additional X-band radar data set which can be used to augment the work already done in this incredibly complicated but significant area of research.

II. Data and Methods

The technique employed closely resembled the work described by Davidson (2003). The experiment was conducted in Southern Monterey Bay, CA where a medium sized Research Vessel with an assumed radar cross section of 2000m2 was observed visually and using an X-Band, phased array radar system with 19.8 Watts power output and 24 dB gain. The bearing and range of the radar site to the vessel location was provided by the ship’s captain for use in initial correlation between the ship and the return on the radar scope. The ship was instrumented with a relative humidity sensor, various air and sea surface temperature sensors and a wind anemometer. Collection of these ‘bulk inputs’ allows for characterization of the environment close to the surface as well as provides the required input parameterization for the Naval Post Graduate School Evaporative Duct Height Model (Frederickson, et al., 2006).

As a first guess at the expected range, the radar horizon was calculated assuming the 4/3 atmospheric approximation of the standard atmosphere (Equation 1). With the ship’s position acquired on the radar, an outbound leg was initiated and two minute segments of continuous sampling of the bearing, range and power density of the return signal were recorded at five minute intervals until the ship approached the calculated radar horizon, which was approximately 24.5 km. As the ship approached the calculated radar horizon, the data was collected continuously until it was longer reflecting sufficient energy to be classified as a target using the radar system. The ship operated inside the 100% probability of detection threshold until it approached 30 km separation. This drastic increase in detection range over the calculated radar horizon for the ship and radar combination indicated that there were ducting effects. Once the ship was no longer detectable, it turned around and a continuous record was made until it was well within the effective range of the radar and the probability of detection was judged to be 100%. This occurred at approximately 27.5km. The difference in range between the bow and stern aspects of the ship was significant and will be discussed later.

The ship continuously sampled the RH%, SST, Tair and wind velocity and direction (Figure 1). These data were used in parameterizing the prediction of evaporative duct height using the NPS model. The primary inputs to AREPS are the characteristics of the sensor and target, as well as the environmental profile. Adequate information was available in order to accurately describe the phased array, X-band radar, however the radar cross-section of the target was not definitively measured and an assumption of 2000 m2 was made. The environmental profile used for input to the AREPS TDA (Figure 2) was based on data gathered over the period of study by the Oakland, CA sounding, the Fort Ord , CA profiler and the organic sensor suite onboard the Pt Sur. The upper air profile was constructed using the output of the Oakland Sounding and the inversion values were derived using the output of the wind profiler and linearly interpolated assuming a constant lapse rate from the base of the inversion to its top. The lower atmospheric profile was derived from ship data and an estimate of the evaporative duct height was calculated using bulk measurements input to the NPS Model.

The data segments collected during the initial outbound runs of the PT Sur were examined in order to correlate the return power density to the range and bearing in hopes of plotting the decay over time and developing a meaningful return curve. The data which were acquired during the transition from 100% POD to 0% POD are unavailable due to an unforeseen system overrun in the size of the data file. This data is not corrupted and is being extracted for future analysis by the vendor who provided the software, Pro Sensing. Qualitative estimates as to the shape of the curve during the fall off from 100% POD to 0% POD indicated a much steeper slope than what was predicted using AREPS. Estimates of the bearing, range and power density of the return were available for ranges up to 23km, but this offered very little insight into the success of AREPS and accuracy of the environmental sampling as 100% POD existed well past 23km.

Although a discussion of model sensitivity to sea surface temperature is presented later, it is important to note that the SST data were collected using 5 distinct methods. The methods ranged from a sophisticated, albeit apparently inaccurate hand-held IR probe to manually measuring a sample of surface water collected using a bucket which was trailed over the side of the ship. The time series over the study period show five very similarly shaped profiles however each has an amazingly different magnitude (Figure 3).

III. Results

The first method employed to determine the expected range of detection was to calculate the radar horizon (Equation 1). This was done by summing the 4/3 earth approximation for the radar horizon of the radar, as well as for the ship, yielding an expected range of 24.5 km. This value is an approximation for the standard atmosphere case neglecting power dispersion. The standard atmosphere case was also run using AREPS and a value of 20 km was determined for 50% probability of detection and 14 km for a 100% probability of detection. The difference in range between the two standard atmosphere cases (calculated horizon and AREPS) is due to the relatively low power output of the radar used, which was 19.8 kW. The calculated value above assumes that adequate power output exists to be limited solely by the earth’s curvature. This was not the case, theoretically, during this study.

The key characteristics of the two different atmospheres evaluated are listed in Table 1. The first is a more restrictive environment from a counter-detection standpoint or can be classified as a more conservative estimate of the ducting effects. The second environment is a more permissive environment from a counter-detection standpoint. For the Pt Sur cross section entered, the 50% and 100% probability of detection ranges were 20.5 km and 15 km for the first environment and 24 km and 16 km for the second environment, respectively. Qualitatively, the falloff curve for each of the three cases appears to be moderately steep (Figure 4) as was expected.

During the range runs made by the Pt Sur, various aspects were presented. The first run was outbound and exposed the largest radar cross section as evidenced by the power density of the return as well at the range to which it was detectable. As the ship approached 30km in range, the radar return dropped from 100% probability of detection to 50% to undetectable in a very short period of time, suggesting that the fall off curve for the phased array, X-Band radar which was used was significantly more steep that that which was predicted. The drop to 0% probability of detection occurred at approximately 30.5 km, which was nearly double the range predicted for even the conservative case using AREPS.

Given the extreme ranges seen in Southern Monterey Bay during the study, the conclusion can be drawn that the extended ranges were due to ducting of some sort. The likelihood of a surface based duct was explored by constructing a plot of the modified refractivity index against time (Figure 5). A negative slope value in the M-Profile would indicate the presence of a surface or elevated duct but the slope of the M-Profile was positive through the column. A logical conclusion about the ranges seen was that the effects of the evaporative duct were being underestimated.

In order to more closely examine the effects of the evaporative ducting, the bulk parameters were input into a model to determine the height of the evaporation duct. The duct height calculated by the model ranged in height from .5 – 5m depending upon the location of the ship and various temperatures input to the model (Figure 6). The variation of inputs will be dicussed later in the paper as options for sensitivity testing, but this result indicates that it is probable that the radar system and craft were operating in the evaporative duct.

IV. Discussion

The results of a comparison between model output and observation indicate that AREPS performed poorly for the case evaluated in Southern Monterey Bay, CA. This conclusion however is somewhat misguided. AREPS is designed to output radar detection ranges based upon a well defined set of user inputs and has been shown to provide incredibly accurate predictions. Although AREPS is capable of handling the effects of evaporative ducting, the height of this duct, and therefore the impact on range prediction, is very difficult to measure. The modified refractivity profile extremely close to the surface is based upon prediction by parameterization using bulk inputs to an evaporative duct model. It is impossible to actually measure the evaporative duct height directly due to environmental contamination by the vessel (Babin, et al., 1997). A typical shape for a modified refractivity profile is normally very negative at the surface and then nearly vertical just off the surface of the water (Figure 7). The vertical orientation of this profile is the primary reason for the extreme difficulty in determining the actual height of the evaporative duct, because the duct height, Z*, is normally identified as the point at which the profile begins to be more positively sloped. For unstable atmospheres, the evaporative duct height is significantly less difficult to determine while for the stable case, Z* is harder to interpret (Figure 7) (Davidson, 2004).

When operations are conducted in environments that are favorable for the formation of such ducts, such as warm air over cool water, conditions similar to those at the Monterey Bay study site, (rapid increase in temperature with height), predictions based on AREPS model output should be discussed as having low confidence. Aside from the difficulty in accurately predicting the height of the duct based on the shape of the model profile, the extreme sensitivity of the model to the SST and RH% inputs adds to the complexity of the problem. This places an extraordinary demand on sensor accuracy as well as on adequate sample spacing for the resolution required. For example, of the five methods used on the ship to measure SST, each had different values with a variation of 2 deg C between the highest and lowest values. The output of three representative sensors was plotted against time to indicate the variability in temperature between them (Figure 3). The model was run using each of these values while holding all other values constant in the environmental profile (Figure 6) and shows the impact of the lack or accuracy and precision on prediction; showing a range of evaporative duct heights predicted by the model ranging from .5 m to 3.5 m. This spread is incredibly significant to detection and counter-detection predictions in that the radar system and freeboard are likely to be fully contained in the upper bound of that value range but would not likely be affected significantly by an evaporative duct only .5 m in thickness.

The results of the under-prediction seen in the Southern Monterey Bay, CA case studied here mirrored the results found by Davidson, et al. (2003) for the X-band radars case. In this case, however, the deviation of observations from predication was more significant than in those previous experiments; the ranges observed were approximately double from what was predicted. The increased effects of evaporative ducting is likely due to the seasonal timing of the study, which was later summer. During the month of July in central CA, the temperature of the air immediately above the surface of the water increases very rapidly with height and humidity, as is normally the case, decreases rapidly with height (Davidson, 2004). There is a significant difference in the actual and predicted ranges between the two studies as well but this is due to the size of the craft used in the experiment. The RV Pt Sur, which closes resembles various MIW ship classes as well as the Patrol Coastal (PC) class of ship in size and therefore radar cross section, is significantly larger than the NSW RHIB boats which were used in the Davidson study. This is difference is not significant, however, since it is the deviation of the observed values from those which were predicted which is of interest to warfighters.

In general, variations in the evaporative duct height are more extreme in littoral regions than over the open ocean (Davidson, 2004). The reason for this extreme variation is the diurnal effect of the sea/land breeze phenomenon. Over the open ocean, the evaporative duct can be very significant however its presence is easier to predict because it is less subject to rapid diurnal changes in temperature and humidity.

A weakness of this study from an operational relevance standpoint, which also exists in the Davidson studies, is the fact that observations were made during daylight hours. This causes heightened evaporative duct effects when compared to what could reasonably be expected during the cycle of darkness. Many MIW operations, and nearly all maritime NSW operations, are conducted during the darkness cycle. In the previous experiments discussed by Davidson (2004), it was shown that the evaporative duct off the coast of CA reaches maximum height at 1600 (local) and is minimal at 0400 (local). This decrease in the effects of evaporative ducting during the darkness cycle should provide some solace to mission planners concerned with the implications of these studies when preparing for operations in the littoral.

A significant issue of concern for naval forces is the increased effects of evaporative ducting in warm water regions of the world. The most active areas of operation exist in warm water regions such as the Arabian Sea, Indian Ocean and in the waters surrounding the maritime continent of Indonesia. NSW and MIW operators are potentially operating under incredibly dangerous assumptions about the potential impact of evaporative ducting. The nature of many of the current threats, normally smaller, loosely organized but very well equipped maritime factions of a larger force, should be cause for concern.

Another weakness in environmental classification is that of sampling spread and resolution. For this case, the profile which was input to AREPS was derived by meshing observations gathered at two land based sensors of known accuracy with the bulk parameters gathered onboard the ship immediately in the vicinity of the ‘target’ (Figure 2). Such in-situ measurements are not an operational reality during military campaigns. Often, METOC officers are forced to rely on information from local public sources, satellite derived inputs or even climatology. It is highly unlikely that such fine sampling and resolution as was used in this study would be available. The seriousness of this is highlighted by the extreme variation in evaporative duct heights when using the various shipboard inputs for SST. It is possible that a tool could be developed which indicated the likelihood of evaporative ducting in situ. Such a device has been discussed by Davidson (2003) and it was suggested that such a sampling unit could be fitted onto all but the smallest of craft. This unit could be used to sample the various bulk parameters needed to predict the presence of an evaporative duct. A simple version of a ducting model could be model run using these in-situ measurements, as well as inputs on enemy sensors, to provide immediate feedback to the operators on the likely threat level. This crude approach is less desirable than being able to accurately predict the presence and effects of evaporative ducting during the planning phase but if it is likely that evaporative ducting exists it would be a useful tool allowing operators to tailor their activities to match the evolving threat level.

V. Conclusion

During this experiment, the effects of evaporative ducting lead to a doubling of radar detection/counter-detection ranges over those that were predicting using AREPS. NSW and MIW units are operating more frequently in warm water, littoral regions where the effects of evaporative ducting are highest. This ducting is very difficult to accurately predict, and is highly sensitive to three critical values, sea surface temperature (SST), air temperature (Tair) and relative humidity (RH%). Accurately measuring these variables is problematic even under the most ideal conditions and nearly impossible in denied environments. The predictions provided by a METOC officer to warfighters should be qualified with an assessment of his confidence in those predictions. The likely presence of evaporative ducting is certainly a significant degradation to the confidence that the officer can expect in his prediction, not due to a limitation of the AREPS system, but rather owing to the lack of observations of suitable accuracy and resolution to handle such complex effects.

METOC Officers using AREPS to develop detection and counter-detection estimates should initially perform a check of the likelihood of evaporative duct presence. Conditions such as extremely stability in a lower atmosphere over warm water lead to formation of these ducts (Davidson, 2004). The suitability of the sensors gathering inputs to the AREPS system should also examined. The purpose of this is to be able to quantify the confidence a forecaster has in the product he is providing. It also allows for a more accurate assessment of the threat level of the environment in which operations are being conducted.

The availability of a suitable, simple sensor system which could be mounted onboard craft operating in close proximity to enemy radar networks would be extremely beneficial in providing in-situ feedback as to a unit’s counter-detectability. The system would be required to measure the bulk parameters of air and sea temperature, relative humidity and wind speed /direction. Davidson, et al. (2003) and Moneymaker (2005) have proposed the funding of such a system to the Commander, Naval Special Warfare for use onboard 11m RHIBs, Mark V Special Operations Craft and NSW High Speed Boats (HSBs). Partial funding has lead to further investigation and the creation of a prototype sensor and associated software, which interface using a PDA.

VII. Future Opportunities

The output of the radar system was analyzed qualitatively during both inbound and outbound runs. A goal of this study was to re-create the fall-off curve, from which the various probability of detection percentages could be determined. This requires a measure of the power density returned to the radar from the contact of interest. Variations in the power density returned would indicate the signal strength. As the intensity decreases, eventually scans would not return any power in the output log of the radar. No set method for evaluating the probability of detection based on the number of sweeps which the radar receives power from the target over time exists. Davidson, et al. (2003) and Moneymaker (2005) proposed counting the sweeps visually and using a pre-determined value of return vs. no- return. Estimating the probability of detection is inherently a flawed process since it is a function of watch-stander vigilance and skill, which are impossible to quantify accurately. This falloff curve was not re-created due to a data problem.

The data needed to re-create this falloff curve is resident on the data logging computer, however it is currently in-accessible to the technicians at NPS. The data are recorded in data blocks and the size of the key data block greatly exceeds the maximum file size that the software can process. Currently, the data is being parsed into more appropriately sized data blocks for future evaluation and construction of this curve. Unfortunately, there is no estimated completion date for this work.

Two excellent research opportunities presented themselves during the data analysis. The first is analyzing the shape of the falloff curve. The output of AREPS indicates a steep curve however the qualitative results of the radar output indicate that the curve seen during this experiment was significantly steeper with the probability of detection dropping from 100% to 0% over a distance of only 1km or so as opposed to approximately 10-12 km predicted by AREPS. The second idea is a correlation of the power density return to sea-swell period. This is an interesting correlation that has significant use operationally. The cyclical shift in radar cross section of the target as it rides up and down on the swell energy can dramatically influence the range of detection. A noticeable pattern did exist in the power density returned from sweep to sweep that indicated that the period of the sea-swell is significant.

Acknowledgements:

This experiment made use of a vehicle mounted X-Band Radar that is operated and maintained by Jeffrey Knorr and Paul Buczynski of the Naval Postgraduate School in Monterey, CA. The use of this radar was critical in the findings presented in this paper. Paul Buczynski aided in the preliminary data collection onboard the radar vehicle.

The idea to construct an environmental profile from land based data collection points in order to overcome the lack of data collected in-situ was made feasible by Peter Guest and Mary Jordan at the Naval Postgraduate School, Monterey, CA. Peter Guest provided guidance as to the most appropriate use of the sensors available and Mary Jordan provided critical data archives needed to compose such a profile.

The crew of the Research Vessel Pt Sur, home ported at Moss Landing, CA was extraordinarily helpful in accommodating the nearshore operations required for observation.

Kenneth Davidson and Paul Frederickson of the Naval Postgraduate School, Monterey, CA, are responsible for the primary study, of which these findings are now a small part. Their guidance and strategy were key in providing a suitable framework with which to conduct such and experiment. The results of their efforts have truly had an impact operationally on the decision making process of METOC officers who have had the advantage of seeing their findings and presentations.

Figures:

[pic]

Figure 1. The output of the ships sensor suit for several key parameters, highlighting the extreme environmental variability over the area of study in southern Monterey Bay, CA. The AREPS model is most sensitive to the inputs of SST and RH%.

[pic]

Figure 2. The composition of the profile used for input in the 2 non-standard atmosphere cases run in AREPS. Such a composition is likely to be the most effective means of characterizing the environment in a denied area where in-situ measurements would not be possible.

[pic]

Figure 3. Despite having access to in-situ measurement of the SST, which would effectively increase the confidence in the output of the AREPS model, there was a great deal of variability in the output of these sensors.

[pic][pic]

[pic]

Figure 4. The three AREPS fall-off curves for the cases analyzed. The first is for the standard environment, the second for the most conservative case and the third for a more permissive environment. AREPS under predicted the 100% probability of detection by nearly 100%

[pic]

Figure 5. The modified refractivity (M) vs. height derived form the 12Z sounding at Oakland, CA site. This profile does not indicate that there is any ducting present but lacks the resolution needed to depict an evaporative duct. The profile is also from a land-based site and does not perfectly match what would be seen over the water.

[pic]

Figure 6. The output of the model for predicting the evaporative duct height. This model was developed at the Naval Post Graduate School and indicates that there is a significant evaporation duct in the area of study.

[pic]

Figure 7. Two typical M-Profiles, the first in a stable atmosphere and the second in an unstable atmosphere. The primary difficulty in predicting the impact of evaporative ducts lies in accurately determining the height of this duct. As can be seen here, this is relatively straightforward for the unstable case but exceedingly difficult for the stable atmosphere case

Tables:

| |Environment 1 |Environment 2 |

|RH% |97 |95 |

|SST (C) |16.0 |16.5 |

|Tair (C) |14 |13.5 |

|Wind spd (m/s) |4 |4 |

Table 1. Table showing the values used for inputs to AREPS for each of the two non-standard cases. The first environment was more conservative and showed longer prediction ranges while the second was less restrictive and showed shorter prediction ranges. Shorter prediction ranges are favorable from a counter-detection standpoint but longer ranges are preferable from a detection standpoint.

Equations:

Equation 1. Calculation of the radar horizon using the 4/3 earth assumption. This value assumes no power losses and only limits the radar range by curvature of the earth. This value was slightly higher than that predicted by AREPS for the standard atmosphere case owing to the power losses from this relatively low power system (19.8 kW)

References:

Babin, S., Young, G., and Carton, J., A New Model of the Oceanic Evaporation

Duct Journal of Applied Meteorology, Volume 36, 1997, 193-204

Davidson, K. L., Jones, F., and Frederickson P., Laboratory and Field Tests of  Instruments for Estimating Combatant Craft

1. Radar Detectability, NPS Report MR-03-011, September 2003, 48 pp.

Davidson, K. L., Assessment of Atmospheric Factors in EM/EO Propagation,  MR4416

& MR3419, Department of Meteorology, Classnotes, Revised 2004

Frederickson, P., Davidson, K., An Operational Bulk Evaporation Duct Model

Submitted to Journal of Applied Meteorology for publication, 2006

Knorr, J., Propagation of Radar Waves, Lecture Notes on Radar

Propagation, Chapter 8, 2006

Kuehn, D., Application of APM and METOC (SMOOS(R)/COAMPS data in estimating small Vessel Detectability, NPS MS Thesis, March 2003, (Secret).

Moneymaker, T., SCI’03 Field Study of METOC Effects in Radar Detection of Low Cross-Section Targets in Coastal Regions, NPS MS Thesis, 86 pp , March 2005, (secret).  

-----------------------

[pic]

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download