Thursday, November 20, 2014

Aquatic Mapping Strategies

Producing professional-quality aquatic maps has never been easier with Lowrance and BioBase mapping technologies, but there are several strategies that can help you optimize your time on the water and produce the best possible map output:

Make Sure Your Transducer is Aligned Correctly
An angled transducer is the most common oversight of users and has been the subject of blogs in 2013 and more recently in 2014 (Figure 1).

Figure 1. Example of a misaligned skimmer transducer and effects on the sonar signal.  For good BioBase outputs, your Lowrance transducer should be aligned parallel with the ground (inset).
A misaligned transducer will result in weak return signals over flat bottoms or slopes where the beam is pointed away from the slope (Figure 1).  In contrast, if the beam is pointed toward the slope, the signal return will be much stronger than normal (Figure 2).  A misaligned transducer can result inaccurate bathymetry, aquatic vegetation, and bottom hardness outputs.

Figure 2.  Example BioBase output where a transom mounted transducer was slanted downward.  In this case, with the boat moving from SW to NE along shore, signal strength is much more diminished on the downslope than on the upslope where the signal is more direct.
Make Sure Your GPS is Aligned With Your Transducer
A second common oversight is large horizontal separation between the GPS receiver (X,Y position) and the transducer (Z position).  This is especially problematic on console-steer boats where the Lowrance Display with an internal GPS may be a good 5 ft (1.5 m) from the transducer.  This means the location of the mapped depth will be 5 ft off in one direction and another 5 ft off in the other direction if the surveyor is doing back-and-forth passes (Figure 3).  A misaligned GPS will result in local imprecision and a "crinkly" looking contour map, but should not affect lake-wide statistics.
Figure 3. Result of a misaligned GPS. In this instance, the surveyor was using the GPS position from the internal antenna in the Lowrance HDS display that was 7 ft away from the transducer where the depths were being recorded.  This positional bias was doubled as a result of the back and forth transect passes (red lines).
BioBase users have two "first party" external GPS antenna solutions which can be mounted right over the transducer and communicate with your Lowrance or Simrad Display.  The Point-1 GPS is a low-cost solution that suits most user needs (Figure 4).  Although the published horizontal accuracy is 5-m, users may expect much better than this, especially in open water.  In 2013, in tests with a differentially corrected Trimble GeoXH 6000, we found differences compared with the internal Lowrance HDS antenna (which uses similar technology as Point-1) to be less than 1-m.  Still, if the user requires more rigorous horizontal positional standards, they can opt for a Simrad HS60 DGPS (Figure 4).  Lowrance and Simrad also support any NMEA 2000 compatible third-party GPS receivers.  Thus with the appropriate NMEA settings and connections, users can stream positions from a wide-range of stand-alone receivers that they may already have as part of other survey work.
Figure 4. Users can align their GPS position with depths by using an external GPS antenna.  Point-1 GPS (left) is the lowest cost and most popular option.  The Simrad HS60 ensures higher reliability of 1m GPS accuracy with its capability of differential correction.  Both antenna are NMEA2000 compatible.

Design and travel transects in a way that maximizes coverage of the features you want to map
A question anyone who desires to make a quality map must address is what is the minimum amount of coverage needed.  BioBase uses the broadband, down-looking 200 kHz signal that collects "point" samples (actually bundled samples from a rapid-firing, 10-20 ping-per-second transducer) directly below your boat (see more about this here here).  Kriging interpolation is a geostatistical way that BioBase employs to predict depth, vegetation, or hardness in unsampled locations.  The more closely spaced the samples (or boat passes), the better the predictions between passes.  You need fewer passes for simple bowl-shaped shallow lakes than convoluted, deep lakes with complex bottom topography.  Same goes for homogeneous vs patchy vegetated or hard bottoms.  Here are some simple transect designs that BioBase users have found successful for creating good bottom maps:
Figure 5.  Parallel to shore design in a flooded reservoir bay.  Note the precision of the internal-GPS track which never crossed structures or shore ensuring confidence in the accuracy of resultant contours.
Figure 6.  Concentric circle transects may be the best data collection approach in small ponds or bays (the pond pictured is 3 acres; 1.2 ha).  In a small boat, kayak, or canoe, the user starts logging as they travel close to shore slowly working their way to the middle of the pond.  This design typically results in smooth bathymetric contours.
Figure 7.  Traveling parallel to the longest shoreline is an effective way of mapping large lakes (the lake pictured is 250 acres; 101 ha). Adding a single trip around shore may enhance the precision and accuracy of nearshore habitats which are often patchy. This approach can be used simultaneously with other lake sampling like aquatic plant point-intercept sampling.  
Figure 8.  Back and forth passes over an experimental plot on Lake Tohopekaliga, Florida USA.  This approach to mapping invasive aquatic plant infestations is much more efficient than using visual cues to know where the edges of aquatic plant beds occur.  We discuss this in more detail here.
Space transects according to the desired level of detail
In most cases, it's unfeasible to completely cover the waterbody you are trying to map. But what area must we cover to produce an accurate and sufficiently precise map?  Users should be prepared to answer the following questions:
  1. What is the size of the waterbody I am trying to map?
  2. Do I want a whole lake map or just a sample area?
  3. How complex are the habitats/bottoms I am trying to map
  4. What level of detail do I need?
  5. What are the consequences of missing some detail if my transects are too wide?
  6. How much time and money do I have available to devote to the mapping project
By default, BioBase produces a 5-m by 5-m square level of detail (grid cell size) for all mapped layers and creates 5-m grid cell predictions to 25 m away from the trip path (i.e., 5 grid cells).  If the adjacent transect is less than 50-m away, the map will be complete.  If the adjacent transect is greater than this distance away, then the area > 25-m is "blanked" and no output is produced.  Users can increase both the grid cell size and buffer "fill" by increasing the buffer in the Trip Reprocessing tab in their BioBase account.  By increasing the buffer, users can "generalize" or "smooth" BioBase outputs.  In some cases smoothing will reduce local precision without sacrificing overall accuracy.  But in cases where significant bottom features are not sampled (e.g., holes, humps, points, patches), then increasing the buffer will reduce accuracy and precision.  This returns us to critically evaluating questions 4 and 5.  Below are some visual examples of how transect spacing and design can affect mapped outputs:

Try a Hybrid Approach
You can cover big water while not sacrificing detail if you employ a hybrid approach that combines wide spaced transects with closer follow up transects in areas of interest.  North Carolina State University showed us how this can be done in North Carolina's largest natural freshwater Lake Waccamaw (8,938 ac, 3,617 ha).  The invasive aquatic plant, Hydrilla, recently invaded Waccamaw and researchers in NC State's Aquatic Weed Science Program (led by Dr. Rob Richardson) needed to know how widespread the plant was to develop a good management plan.  They started with running 300 m transects on a couple of boats, while also collecting vegetation point samples.  This took about a week to finish the field work and a map with 60-m grid cells was processed and merged in BioBase in a matter of hours (Figure 9).
Figure 9.  300-m transects (red lines) on Lake Waccamaw overlain onto a submerged vegetation map processed by BioBase.  Red represents vegetation growth close to the surface, green is low-growing vegetation.  Blue areas have no vegetation growth.  Grid cells were 60 m.
Physical plant species samples suggested that most hydrilla growth occurred near the boat launch in the NW part of the lake which was not surprising since most new species introductions originate at boat launches.  Accordingly, NC State researchers required some more detailed information about the plant beds in this area and implemented a more intensive survey in the launch area.
Figure 10. 50-m transects (red lines) on Lake Waccamaw recorded over an area near the boat launch with relatively high hydrilla cover. Grid cells are 10-m
When we zoom in and look at the map outputs of the alternative mapping strategies we can make some informed,visual conclusions about alternative mapping strategies (Figure 11).

Figure 11. Vegetation biovolume maps of the NW corner of Lake Waccamaw processed by BioBase and converted to a raster in GIS.  The map on the left was created with 50-m transects; the map on the right was created by 300-m transects.
The first striking comparison between the 50-m transect map and 300-m transect map is the difference in the detail, patchiness, and highs and lows of vegetation Biovolume.  However, when you look at the statistics, on the whole, the percent area that has vegetation present and the overall vegetation height (expressed as avg percent biovolume) doesn't look too much different.  This is a visual, geospatial representation of the difference between accuracy and precision.  Both are accurate results and show the same general trends.  However, the map on the left is more precise due to the higher number and closer spacing of transects than the map on the right where the map was generated from maybe one or two transects.

Now you are empowered!
This blog was meant to cover mapping strategies A to Z quickly and get you on your way to creating high quality BioBase outputs with your Lowrance or Simrad Sonar and Chartplotter.  If this brings to mind questions or creative ways you've navigated these issues, please comment or sign up and post a forum discussion at http://aquaticmapping.prophpbb.com/.  We are always impressed with the innovative mapping solutions of BioBase user community.  This community had presented these great examples that help newcomers to the technology get up and running quickly!


1 comment:

  1. Users looking for additional external GPS solutions with Pitch/Roll/Tilt compensation might consider the Simrad HS60 that is compatible with Lowrance via NMEA2000. Read about it here - http://www.simrad-yachting.com/en-US/Products/Compasses/HS60-GPS-Compass-en-us.aspx

    ReplyDelete

Please share your thoughts about this topic with us