Wetland Wednesday
September 2, 2020
The US Army Corps of Engineers (USACE) released a draft version of its Antecedent Precipitation Tool (APT) on July 8, 2020. The tool is a desktop-based web interface written in Python by Jason Deters of the Corps. It is Windows compatible, but it does not appear to support Apple. It is also not mobile ready.
The purpose of the APT is to simplify the rainfall analysis now required for a wetland delineation as described in the new Navigable Waters Protection Rule (NWPR). The NWPR necessitates that wetland delineations should only be conducted when rainfall amounts on or before the delineation are considered normal for the season. This translates to rainfall amounts occurring within the 30th and 70th percentiles of the daily mean. Below 30% is considered too dry and above 70% is conserved too wet to conduct a wetland delineation.
There are a number of rainfall analysis tools in use currently. However, there are none that are explicitly designed for wetland assessment proposes. This new USACE APT is designed specifically for wetland delineation use.
It is important to note that this new APT is in beta form. It has not been formally released by the USACE and it is only available to the public via GitHub.
There are a lot of positives about this new tool. The first of course is that it is free! Second it does a really nice job of plotting out graphical illustrations of the rainfall trends. The tool automatically prepares a spreadsheet of the entire data review and produces daily graphs in PDF format. Both are easy to read and there is a lot of in app help and help guides back on GitHub.
The tool starts up in a command box and an interconnected graphical user interface (GUI). It has a bit of a DOS look but I would rather see what it is doing rather than hiding it and just using the GUI. So, I will also count the DOS look as a positive.
To run the tool the user simply enters the latitude and longitude of the project site and either a single date to review or a date range. That is it. The program does the rest. It will pull down recent rainfall data and compare the current data to historic data via another web tool called WebWimp. Once it is done you have a nice graph and spreadsheet that you can dive into.
There are some negatives to this tool. The first is the ability to install the tool onto your computer. Anti-virus programs and many server and local cyber security tools will prevent it from installing and running. These all must be turned off. Good luck getting your IT staff to let you do that. I would recommend buying a cheap laptop and using it on your company’s guest network or other unsecured network (Starbuck’s?). Just do not install any security software. Windows will also put up some warnings, but you can bypass these once it gets done with the “are you sure” messages.
From a wetland science perspective, I only have one issue with the program. The dependence on WebWimp for historic data is a bit of a concern. WebWimp is a web-based system hosted and developed by the University of Delaware. Its last update was around 2009. Being over 10 years old is not the concern. We are looking at historic data after all. The issue degree of accuracy of the data relative to the project site.
WebWimp is basically an online portal to an old database. The user enters the site latitude and longitude and it almost instantly produces a wet, dry and normal year result. The USACE APT does this through a direct database call between the two systems.
WebWimp is not doing any calculations. It is too quick. It is just doing a lookup. The publisher indicates that it is accurate to half a degree of latitude and longitude which is 30 plus miles in either direction. I suspect that there is also a lot of extrapolation between even these points.
At issue is that the APT is running a very accurate local analysis of rainfall data and then comparing it to very generalized historic data. The problem that results from this is that the historic data does not account for on-site conditions like rain shadows and other localized weather phenomenon that are ongoing.
There are other historic rainfall analysis tolls that would use the same APT precipitation data set. WetBud for example will take 30 plus years of weather data and calculate wet, dry and normal years based upon local weather data. I suspect that this could be a feature added into the APT in a future version fairly easily.
Overall, I really like this new APT. I would suggest using WetBud to generate your own historic data calculations and compare them to what the APT puts out. The math to adjust is based upon “real” data and would simply be accounting for the 30th and 70th percentile range of the calculated data. In short, I give this latest USACE effort two thumbs up!