In order to measure the water level in a rain water cistern, I am using a Grasshopper LoRa/LoRaWAN Development Board and a MaxBotix Ultrasonic sensor. The main question is: How long will a 2000mAh LiPo battery last if the current water level is sent every 30 minutes via TTN?
The complete project setup is described here on hackster.io
Is there any chance you could have a stab at the answer yourself - you will learn a lot and more people will join in to refine the answer. Otherwise one of us could just dump a value in your lap without any real investigation.
Think about the current consumption of the different phases of the normal cycle. Don’t worry about the real world figures at first, those can be refined once you’ve got a working method.
Google Sheets may well be your friend.
Also roping in @onehorse as the developer/owner of the grasshopper to help with the device details (note also did an ultrasonic (bin) monitor - using an STM ToF sensor IIRC) and to add some background and colour 
You can rely on the numbers mentioned on the Tindie product page but the best way to determine average power usage is to use a fully charged 100 mAH LiPo battery and let your application run until it stops. Then you can directly determine average power.
I would recommend for most applications to 1) cut the power on led trace to disable, 2) run at 4 MHz CPU speed and 3) make use of STOP mode. In my experience, the Grasshopper sleep current is 2.1 uA in STOP mode and sending LoRaWAN Tx every 10 seconds costs ~40 uA average power. YMMV.
Which we would never do in real life unless we have put a 50Ω resistor in place of an antenna and put it in a metal box for the 100mAh battery test.
It shouldn’t change based on how often it sends, but should change depending on DR.
Yes, this should have said every ten minutes. This might still be too much for TTN but I do this regularly in deployment.
I’ve got one on my desk sending 6 bytes every 3 minutes on DR5 for a sustained period and then revert to a more relaxed cycle - well inside the FUP. I’m trying to determine if/how the battery bounces back after a mild hammering.
Often your total charge per day is dominated by the idle current. Even though it is much smaller than for example the transmit current, idle current is spent 24 hours per day. Transmit current is spent only 30 seconds per day, or about 0.035% of the time.