Hello All, I was wondering if anyone out there is well versed in the USGS data. I'm wondering if there is a way to graph each year (jan 1st- dec 31st) and then lay multiple years on top of each other to get a comparative idea about flow windows for the spring? Basically I'm looking for a visual representation of the peak flows.

Is this possible or do I just need to look at each year on a case by case basis? Is there another resource that would be a better tool for this? I'm trying to avoid making too much work so if it means too much technical knowledge I'll just go the old route of having multiple tabs open in one window.

Any advice would be great.

Pray for snow.

I don't think the USGS site has a way for you to overlay multiple years on a single graph. But you can download the data for each year as a TSV file from the USGS (they call it RDB format), and then import that data into Excel. Once it's in there, you can graph it. The process is a little tedious and time-consuming so if you're going to do more than a few gauges, you may want to automate it somehow.

   The other, much easier option, is to use Dreamflows - if they have the gauges you're looking for. Once you click through to the graph on their site, most gauges have a graph that shows the previous 3 years on a single graph. Here's the link to the Dreamflows page for Washington: http://dreamflows.com/flows.php?page=prod&zone=panw&form=norm&mark=All#Washington_Rivers
   For example, here's their graph of Ernies for the past 3 years: http://www.dreamflows.com/graphs/yir.280.php

   If you want/need the USGS data, here's an overview of the process:
   Since you're looking for yearly data, it's probably better to download daily values instead of 15-minute values. For example, here's the USGS page for daily values for Ernies: http://nwis.waterdata.usgs.gov/nwis/dv?cb_00060=on&format=gif_default&period=&begin_date=2013-12-01&end_date=2013-12-17&site_no=12142000&referred_module=sw

   If you select only discharge data for this year (2013) and select "tab-separated", you get this page: http://nwis.waterdata.usgs.gov/nwis/dv?cb_00060=on&format=rdb&period=&begin_date=2013-01-01&end_date=2013-12-17&site_no=12142000&referred_module=sw

   You can save that page as a text file (.txt) and then use Excel's Data Import from Text feature to pull it into a spreadsheet. There's a wizard that walks you through the import so you can see what you're going to get before you actually import it. But the only 2 columns you probably need are the datetime and the flow.

   Then repeat for each year you want, and add the levels for each year as a new column (with the rows/dates matched). Once you've got all the years you want, chart all the data with the days as the x-axis and the flows for each year as separate colors on the y-axis.

If you go the USGS/Excel route, post whatever charts you come up with!

I'll be interested to see if someone can chime in with a more straight forward approach.. This is my best guess. its process intensive, confusing to write, and could offer no new information. Here is how I would get at what I you want.

1. On your gage page of choice click "show all data available for this site"

2. Choose daily statistics

3. Make sure discharge is selected and enter your date range of interest in the format it asks for. You probably want the whole period of record unless something has changed in the basin that attenuates melt/runoff.

4. Here is where it gets tedious. You'll want to select "Table of XX percentile", select the percentile you want, submit and copy out the data into excel. I'd say start with the min, 25, 50 (median), and 75. Add in max if you want for the peaks.

A little bit on the percentiles, I personally feel its confusing and backwards. Selecting the 25th percentile is like saying there is a 75% the flow will be at least XX CFS. 90% is like saying there's a 10% chance the flow will be at least XX.

5. Once you've copied your tables into excel you will need to rearrange the data and group the months together. (Ex. Arrange all the Aprils next to each other). The reason I would do this is because it will be much easier to look at just 1 month per graph you create.

6. Once you have all your "Aprils" next to each other, highlight it all, and create a XY scatter.

The product will be a graph for the 'month of interest'  that shows the historic probability of a flow occurrence for the day of record. You will have to do that for each month. Its not the fastest but once you get going its not toooo big a deal.


I'll note I don't know how this translates to flow windows as you mentioned.