Friday, November 26, 2021

Sub-millibar Pressure Patterns, Part 2.

In a recent note (Atmospheric Pressure: Look Close to See its Microscopic Pulse) I presented high-resolution results of three barometer sensors in three phones here in the office using our Marine Barograph app. This app is well designed to measure and record small changes in the pressure, and indeed we saw small oscillations in the pressure that were observed on all three, presumably independent devices. 

The question was raised as to whether there could be any ambient interference with the pressure or electronic circuits inside the room that were responsible for these oscillations, so we decided to move the three phones outside and then put them into a metal box, that serves as a Faraday cage that blocks out wireless interferences. Also these were now running on batteries not connected to chargers as the first measurements were, which does indirectly link them via the ground line at least.

This new sophisticated (!) set up is shown below.

We let the units collect data for 40 min or so, and then we used the app's export .csv option on the 30-min files for each, which includes the pressure every second. These data where then imported to Excel, trimmed to be the same start and stop times, and again shifted one of them that did not have its offset turned on, as earlier.  The results are as shown below, which we annotated with the colors.

We learn a lot from this result. First, it looks like what we saw earlier measured indoors, namely about 0.1 mb oscillations at about a 70-sec period was somehow unique to being indoors and connected via the ground connection. All three showed that pattern on some level and they were more or less in sync. I believe now that this was an electronic cross talk between the green one above and the other two. That sensor circuit seems to be defective on this very low background level, and that oscillation is almost certainly artificial.

Second and foremost, however, is that within the clean environment of these new measurements, there is still a real, physical variation of the pressure on a sub-millibar level that is clearly detectable with this $15 app.  The scale of the variation shown is really small, with peaks showing up on the 0.1 mb level with periods of about 5 minutes.

It could be pulses in the actual atmospheric pressure pattern or it could be disturbances to that pattern from other sources. We see the pressure is clearly rising over this period (~0.6 mb in 26 min), and we see the rise rate increasing as well.  This is all good tactical information, totally unrelated and not dependent on any ripples in the curve. The navigator at sea would care about this early detection of a pressure rise, but does not care at all about the ripples on the curve.

On the other hand, an atmospheric scientist might care... or maybe more likely, an engineer entrepreneur might see this and wonder if that is detecting the passing of a jet plane that I cannot see or hear?  Would the military care about such a device if really works this way?  Next time you are waiting at the airport with an iPhone and this app, you can collect the data to test this WAG proposal!

We know that this app can detect wind gusts, but these are much larger spikes in the pressure on a much shorter time frame. The 5 minute period is intriguing.

In any event, it remains promising that we can purchase an app for so little that has all the functionality to be an actual scientific instrument. Crowd sourced measurements of pressure might lead to new discoveries, or at least maybe contribute to improved weather forecasting.  Mariners have no excuse at all these days to go to sea without a high precision barometer in their pocket... or as we recommend, get an old one and velcro it to the bulkhead to be aware of the shape of the pressure changes at all times.

Wednesday, November 24, 2021

Depths, Contours, Soundings, and Groundings in ENC Navigation

Electronic navigational charts (ENC) provide a vast array of safety features not available using raster navigational charts (RNC), the echart equivalents of paper charts. These notes explain how charted depth information can be used in an electronic charting system (ECS), and in particular how qtVlm is well designed to study and practice these features using its built in NMEA simulator for power or sailing vessels. It is a good representative of any high-featured ECS. We use it in our forthcoming course on electronic chart navigation.

The subject is tied to how ENC present depth contours and soundings. Indeed, the proper use of depth contours is a key to good work with ENC. To be more precise, we are not really talking about properties of the ENC themselves, but rather, what ECS such as qtVlm can do with the specified format of the ENC sounding data in compliance with recommendations of the IHO and IMO

These two organizations call for the definition and use of three specific depth contours, and all modern ECS use these conventions. The user defines in the program preferences the values to be used for each of these contours, keeping in mind that they can be changed at any time to match the location and navigational goals of the vessel, as well as the scale of charts available. The three are Shallow water contour, Safety contour, and Deep water contour.

Shallow water contour

The shallow water contour is defined as marking that depth where the vessel will certainly go aground. It would typically be the draft of the vessel, perhaps modified by the range of the local tide. The working assumption is if you cross that contour you go aground, but there are important differences between the one requested, the one displayed, and the one that triggers alarms, as discussed below.

Safety contour

This is the most important of the three. It is intended to mark the boundary of guaranteed safe water for the vessel, meaning if the vessel remains outside of the safety contour they have no worries about grounding. This contour will be highlighted on the chart with a thicker, darker contour line, and its depth value is used to identify any isolated obstruction outside of the safety contour with the prominent isolated danger symbol (magenta circle with a transparent bold X inside) if its sounding is less than the safety contour depth you requested.

The safety contour will also be marked by two different water colors on either side, which will be true regardless of whether the mariner chooses a 4-color water pattern or a 2-color water pattern. The IHO and IMO also require that the COG predictor line (or anti-grounding cone) also trigger an alarm if the predicted vessel position crosses the safety contour.

Generally the safety contour would be chosen to be the draft of the vessel plus the user's choice of a safety margin below the keel, plus a correction for negative tides in the region. Ships and other large vessels also add a correction for maximum squat they might expect.

Safety depth

The goal of most routing would be to stay outside of the safety contour, but in some cases this must be crossed, and when doing so, special caution is required. One ENC display feature that assists with this is a fourth navigator input to the navigation program called the Safety depth.  The value of the safety depth is used to determine the color of the soundings.  If a sounding is equal to or less than the Safety depth, it should be printed in black text; if the sounding is deeper than the Safety depth it is printed in a gray color. Thus when sailing inside of the safety contour it can be easier to tell the deeper parts from the shallower. The parts deeper than the safety contour have gray soundings; the parts shallower, have black soundings.

The safety depth value is also used to change the background color of the important generic hazard symbols from transparent, when their soundings are deeper than the safety depth, to blue, when their soundings are equal to or less than the safety depth. See Role of the Safety Depth in ENC Display.

Deep water contour

This contour choice does not directly affect safe navigation and can be used as best suits the navigator. In a 4-color system it defines the outermost color boundary. We see several formulaic guides in online articles for choosing this, but they just generate a number with no specific meaning. We have suggested two practical applications, namely mark the 100-fathom contour on a coastal voyage to avoid the sometimes confused seas and currents along the edge of the continental shelf, or set it to your preferred anchoring depth to help look for anchorages ahead. Or maybe just choose the next deepest contour beyond the active safety contour to make a progressive color pattern to the water depth.

This contour is not important, other than it should be deeper than the safety contour, which in turn should be deeper than the shallow water contour. Generally the safety depth would be equal to the safety contour value or between that and the shallow water go-aground contour.

Setting ENC depth contours: The good, the bad, and the ugly

The good part is the contours we can choose, their clear presentation, their intentions, and their interaction with the ECS are as described above, which in principle provide valuable navigation aids. We have no such options using other chart forms.

The bad part is, selecting the contours we want to use is not as easy as we might guess. In all ECS, we are given the opportunity to type in the digital values we want for each of these contours (the requested values), but rarely will we actually get what we want drawn out on the chart (the displayed values). The problem is the only contours that can be displayed are ones that are already coded into the specific ENC we are using—and the only contours on the ENC are those that are on the RNC (paper chart) it was based upon. Thus if we want and ask for a safety contour of 20 ft on a US chart, we won't get it, because no US chart has such a contour on it.  

When this happens, ECS are instructed (by IHO S-52) to find the next deepest contour included in the chart at hand, and display that one. Requesting a 20-ft safety contour, will likely lead to a 30-ft safety contour, or on some US ENC there is also a 24-ft contour (more rarely) and it would take that one. Table 1 shows the contours available on ENC.

It is beneficial to know the gist of this table, determined largely by fathom charts, but even knowing the possible contours is not the end of the story. First, not all US charts have all these contours. You might have two charts next to each other with different contours. Samples (not next to each other) are shown below using a nice display feature of qtVlm that tells us all the contours in a specific chart in view, with the active safety contour underlined.

Beyond knowing what the options are, which is easily solved with this display, there is still a nuance in asking for it that can be traced to NOAA's policy of truncating the metric conversion of depths when creating the ENC based on the existing paper chart. The details are given in our book Introduction to Electronic Chart Navigation

In short, if you want a specific US contour and you have the ECS display set to feet, you must ask for something a foot or so lower than you want. If you ask, for example, for 18 ft, it will miss that one completely because it is in there as 17.7 ft. Ask for 16 to be sure to get 18, and so on.

Knowing how this works, one safe method is to look at the contours on the chart in use (viewed as feet  or meters) and cursor pick the contour you want as a safety contour to read its value, which in this case might well be reported back as 18 ft (or 5.4 m), but then ask for 16 ft, then look at the chart to confirm that the one you want is now bold with different colors on either side. Set to meters display, asking for 5.4 m will likely get you the one you want.

This extra concern called for in contour selection will all go away in a couple years, depending on where you are sailing. NOAA has an ongoing program of Rescheming all ENC and one of the many benefits of that program are new and consistent metric contours, as shown below.  Some areas are already completed, but others are a couple years out, as shown below.

The same considerations on setting the Safety contour digitally also applies to the Shallow Water (go aground) contour and the Deep Water (use as you please) contours.

An example of entering requested contours and seeing the consequences. Notice that soundings equal to or deeper than the Safety depth are in gray; others in black.

Subtleties in the use of the Safety Contour

The IHO and IMO require that ship navigation with ENC must offer the user the option to set alerts and alarms that will go off when the vessel approaches or crosses the safety contour. Most ECS that have this capability have also adopted this as an optional set up for the mariner. Thus they can turn on a COG predictor or anti-grounding cone and it will change color or set off some other visual or audio alert when it crosses the safety contour. You can choose to look ahead for any time interval, minutes or hours. It will also trigger when crossing any isolated danger (rock, wreck, or obstruction) outside of the safety contour whose sounding is less than the depth of the safety contour, or if no sounding is given.

This is a fine concept, but referring back to the notion of a sometimes bad aspect of the use of a safety contour, consider the case when there is not a contour in the chart that is close to what you want, and the next one deeper is quite a bit deeper than you want. When that happens, your otherwise useful alarm is now a nuisance. It is going off, for example, when you are crossing a 60-ft contour, when you have no danger at all above 12 ft, which is what you ask for.

One solution is just give up the audio alarm feature till the chart contours are more favorable, but there are other solutions that help clarify the situation.  qtVlm, for example, has implemented the system of Default Safety Contours recommended by Professor Adam Weintrit in his book The Electronic Chart Display and Information System (ECDIS): An Operational Handbook. In that system, if you request a safety contour that is 10 m (33 ft) or less and the only available safety contour in the chart is more than 67% deeper than you requested, then that selected contour is considered a default safety contour and it is marked with a less prominent line (lighter gray) than used for a closer safety contour.  It is still clearly different from other charted contours, but also different from that of a safety contour that is closer to what you requested. If the requested safety contour is deeper than 10 m, the default contour coloring is used if it is more than 33% deeper than you requested.

This is a seamless, behind the scenes enhancement to the presentation of the safety contour that reminds users when the safety contour in effect is not close to what was requested, and yet all other interactions with it remain unchanged. This is especially valuable when there is a mismatch of contours on adjacent or overlapping charts, as shown below.

Here we see a default safety contour in lighter gray in the bottom chart compared to the traditional ENC safety contour in the top chart. In both cases, we asked for 19 ft, and the top one had a contour at 30 ft, which was close enough to what we wanted to treat it as a traditional safety contour, but when viewing the bottom chart, its closest contour was 60 ft, so it gets relegated to a default safety contour status, which stands out from the other depth contours on the chart and also from the closer traditional safety contour that was closer to what was requested.

We also see here the danger cone alerts crossing the safety contour (orange) and the effective shallow water contour (red)—see related notes below. The safety contours are the ones in effect on the active charts. The orange alerts marked with yellow asterisks (discussed below) are going off because of the mismatch in contours; the COG predictor is detecting the safety contour from the adjacent chart, which can be seen on the lower chart.

Notice that soundings equal to or deeper than the Safety depth are in gray; others in black, which is a big aid when sailing within the safety contour.

Isolated danger symbol and the safety contour

Another nuance of the official safety contour procedures relates to the display of the isolated danger symbol for rocks, wrecks, and obstructions that are always underwater and lying outside of the displayed safety contour. These object symbols will change to the isolated danger symbol (prominent pink Philipps-head) whenever their known sounding is less than the depth contour that you requested, not the one that is actually displayed. Thus there is some virtue to requesting the safety contour that you really want (even though a deeper one might get displayed) and also setting the safety depth to this same depth. That way all the soundings will be black in waters shallower than that and any underwater hazard shallower than that will be danger symbols.

The ugly bits

What I had in mind as not a very pretty part of using the safety contour is the fact just mentioned that in the present state of US ENC, meaning before the rescheming is completed, we have cases where adjacent or overlapping ENC do not have matching depth contours. The above image is a good example. Different scales do not matter, but when the inherent contours differ it means the safety contour in effect can change, depending on the chart in view. The ECS is always choosing the next deeper one from what you requested if it does not find what you want, but that answer can change as you sail from one chart to the next, or you change the zoom level to change the active chart. Although this happens on other charts as well, the example shown here is an unusual exaggeration of this issue.

This problem is mitigated quite a bit with the use of the Default Safety Contour system, which alerts us to the presence of more than one safety contour, but it can still lead to unusual danger cone displays (the two with yellow asterisks shown above), because the ECS knows all the contours present on all of the loaded charts, whether or not it shows on the chart in view. It has to know that to trigger the alerts. Indeed, a great value of the ENC is you can load the ENC, and then navigate on an RNC and the alerts and alarms computed from the ENC will still be in effect. In this case it looks like the alarms are coming from the RNC, but in fact they come from the ENC below these that you cannot see.

Thus if you see your danger cone go orange in what looks like the middle of Nowhere, you can know that you are crossing the safety contour as defined on a chart you cannot see. In short, there is no safety issue here, we are just finding one of those places where the depth contours on overlapping charts are not the same. There are not many places this can happen, and in a year or so there should be none.

Danger cone alarms are also triggered by isolated dangers shallower than the active safety contour depth, so this behavior also changes in cases like this one.

Depth simulation in qtVlm

For navigation training and practice, the NMEA simulator included in qtVlm is an invaluable aid. There are several videos on the use of this tool, but for now I want to just summarize a new feature directly related to the topic at hand. That is, it will now simulate depth readings as your vessel moves across the chart. It can do this because no matter where you are on the chart you are between two depth contours (d1 and d2), and these define an ENC object called Depth area (DEPARE), which has two attributes, d1 and d2.  qtVlm simulates what your depth sounder would read by reporting back a depth of (d1+d2)/2 at all times, and you can display this in the depth meter (sounder) found in the instruments selection.

This is clearly a rough approximation, which appears as a bar graph in the histogram, with steps occurring as you cross a contour, but nevertheless this simulation can be used as a way to study depth sounding navigation as shown below.

If you had just the depth sounder trace shown above with no GPS working, you could find your position on this chart by matching the trace to your known heading and speed. This can be practiced with the simulator.

The simulator reads the depth below the location of the GPS on your simulated vessel that you specify in the boat dimensions tab (± 10 meters). These dimensions are entered in the same way they are specified if you were broadcasting your AIS location. Likewise when we get a ship's AIS signal, this is the way we learn where its GPS is located.

This shows the set up for a real size vessel icon with the GPS assigned to the starboard quarter, marked by a small red dot. The yellow dot forward, is one third of the LOA, which marks the rotation point of the vessel when turning in the simulator. The red ring with radius 10 m is effectively the simulated depth sounder profile on the sea bed. 

Going aground in the simulator

The simulator will go aground and stop the simulation whenever you cross the effective shallow water contour.  Recall from above, that if the shallow contour you request is not there, then the next deepest will be chosen for the water color change.  Also, if the exact contour you requested is not there, the shallow alarm does not go off at that displayed shallow contour and you do not go aground at that contour, but rather at the next shallower contour. In other words, if there are 5, 10 and 15m contours, and you ask for shallow at 7, then the 10 will be displayed, but you do not go aground or get the obstruction alarm until you cross the 5. 

You can test this behavior with the Rule tool. Just turn it on and extend it over the various contours or objects to see the color changes and notices. You will go aground whenever you detect a red obstacle warning.

If there are no shallow contours where you approach land, you go aground at the green foreshore, and with no foreshore, you go aground on the tan land.  With no ENC charts loaded, you go around at the boundary of the base map. There is a setting in Preferences that affects this choice, which is best set to the ENC chart borders. 

The grounding applies not just during simulation, but it is also taken into consideration when qtVlm  computes optimum routes, power or sail. Needless to say, it will not propose a route that takes you aground!  In this regard, the details can matter. If you do not have real dimensions assigned to your boat, then the grounding occurs when the vessel's GPS location crosses the shallow water contour or coastlines as noted above, but if you assign dimensions to your boat, then it is the rectangular profile of the boat that triggers the grounding. If your boat has, say, a beam of (C+D) = 4 meters, and your GPS is on the centerline (C=D), then you could drift abeam and go aground 2 m before the GPS gets there.

When running the COG predictor or danger cone, you will get an orange warning when the predicted position (which depends on the look-ahead time you have selected) crosses the safety contour, and you will get a red warning if it crosses the shallow-water contour.  If you have the audible alarms turned on with the danger cone activated, then these warnings trigger a graphic sign alarm and a sound. 

These warnings and alarms will also get triggered by crossing any isolated danger symbol or indeed any obstruction in any location whose sounding is less than the active shallow-water contour. The danger cone is the best way to afford this safety and to practice its function. The COG predictor line alone is less likely to hit the point source of an obstruction. 

You can investigate which alarms go off and when using the ruler tool, because it acts like a portable COG predictor and will go orange crossing the safety contour and red crossing the shallow water contour or any obstruction shallower than the shallow water contour.  You can sweep the Ruler tool line across an obstruction to see that alert. Orange alerts are announced "danger detected"; red alerts are announced "obstacle detected." Both can be used to trigger an extra visual and audio alarm, but the latter must be turned on manually first.

You can get realistic practice on grounding with dimensions assigned to your vessel, because then you will indeed go aground when that vessel outline passes over an obstruction whose sounding is less than the active shallow water contour. However, with the danger cone activated, you should not have this happen because it will warn you. That is why we are practicing this! Turn on a notable tidal current (Grib configuration / Corrections)  to see how this works when you are not going the way you are headed. 

To have the alerts, groundings, and routing restrictions work properly in the program, we must have the coastline detection set correctly, which is done on the ENC setup page.

If this is shut off, then the coastlines are defined by the borders in the base map, which are not nearly as accurate, and we would sail over or route over other shoalings as well. Likewise, assigning these to an ENC without it being loaded has the same effect.

Monday, November 22, 2021

Atmospheric Pressure: Look Close to See its Microscopic Pulse

We had an inquiry today about whether or not our Marine Barograph app would read and record pressures accurate to a few tenths of a mb on a per second basis and plot then showing sub mb precision. The answer is yes it will, but the question caught my attention. I have known for a long time that pressure oscillates on a small scale at all times, but this never effects our weather work, which looks for solid trends of over a few tenths of a mb—in a sense for just this reason, namely we do not want small oscillations interfering with our reading, no matter where they come from.

When we look at a plot of the pressure on a fine scale we see these oscillations, but when you look at just one plot the tendency is to consider this electronic noise. You might think you are just looking at the limits of the sensor and its circuits.  But that is actually not the case at all. These oscillations are real.

Below we see three old cell phones with our app loaded into them, converting them in a sense to about a $800 electronic barograph. Any iPhone model 6 or newer includes a pressure sensor—you may have one sitting around in a drawer.

The dip in the top unit on the left side of the graph of about 0.2 mb marks the time when, upon receiving the inquiry about the units, I picked this one up and raised it up to the ceiling, a distance of about 5 ft. This was to check how the scale responded to such a change.

The middle one did not have its sensor offset turned on so the value is off some, but this does not affect the topic at hand.  

Data are stored every second for 30 min, and then stored in different time bins for plotting longer periods, which is described in the help file.  You can export any of these time bins, which I did for each one covering the past 30 min.  We can see the plot of these in the phones themselves, and indeed pinch zoom in to see all details, but to make the point at hand, I exported them and then plotted them in Excel, as shown below.

Here we see all three plots; I shifted the gray one up to compare them, to compensate for its lack of calibration offset being turned on.  This whole note was pretty spontaneous!

The important thing to note here is these are three completely independent sensors, being run by three completely independent little computers, which are powered by batteries. I cannot imagine any common electrical properties of any of them—although the batteries were all on chargers at the time that does in  a sense link them.

In short, these oscillations in the pressure appear to be real variations in the physical pressure of this room. I do not know what causes it. Back when i was working in a laboratory, I would likely have stopped to figure out why, but now we just note this and move back to what we are now working on—a new course in electronic chart navigation.

Hallways in condo buildings are often kept at positive pressure to keep cooking smells indoors. Navy warships and first responder vessels keep their wheelhouse in relatively high positive pressure to protect from toxic gases, so these things could be studied with our app. Barometers in these ships have to be in sealed housings that have tubing leading outside to read atmospheric pressure.

There is some element of pattern present as very roughly outlined below, but this would take a serious mathematical analysis to understand.

This rough outline indicates that the oscillations have an amplitude of about 0.05 mb with a period of just over 1 minute. It is not clear what is causing this, but it seems it must be a real physical phenomena and not some artificial electronic anomaly—but that is not 100% clear. We have to rule out any link via the battery charger and investigate any way a wireless signal might induce this effect.

We will take these outside tomorrow for 15 minutes or so to see if that  makes a difference. They will also be on batteries without chargers for that and I will put each one in a metal can (Faraday cage.)  I don't think it will matter. I think we are seeing the pulse of the atmosphere here!

And I want to stress that this is no reflection on the actual accuracy of the pressure itself. With a good offset inserted, the reading is likely accurate to within a few tenths of a mb, which we know from our own calibration tank and comparison with other instruments, but we can't make comparisons any closer than that. But even though the accuracy on the full pressure readouts are not in this 0.01 mb range,  the oscillations themselves are likely real. That is the point of this note.

When we first started working on electronic barometers we also reported a very interesting physical affect of a small short pressure bump under a squall, which is indeed atmospheric science, but we have not learned more on this since then, although you see in that article a reference to an even earlier event we reported—we were not lucky enough to have our new app in these early days.  Maybe now that our app is available, more folks will study such events.

In short, a simple tool like our $15 marine barograph app is a powerful way to look into details of the atmosphere, besides its great value in helping shape a fast, safe course across the ocean... or predict when a front will cross our local lake

Let's finish with a bit more perspective on this. Here is the dial of a high precision Fisher aneroid barometer, which has the pressure scale marked off in 0.5 mb intervals.


Note added Nov 26.  We have now done the outdoor test and the oscillations noted above are likely artificial, but we did indeed see what are more likely real sub-mb oscillations. See Sub-millibar Pressure Patterns, Part 2.