In 2018, Dr. Richard Koehler presented a Golden Software webinar on how the Surfer package can be used to detect patterns of daily, monthly and seasonal streamflow on the Colorado River. He demonstrated a new way of plotting historical flows using a raster map technique to visualize time-series data.
The traditional way of plotting streamflow at a given point on a river is with a 2D line graph where time is on the X axis and flow is on the Y axis. This may work fine for a limited period, such as a year, but when multiple years of line graphs are plotted on top of each other to visualize river activity, the result is often a confusing 'spaghetti plot' where trends and patterns are hard to pick out. It becomes nearly impossible to compare the values of a single year to any other year due to the jumble of lines.
Koehler, CEO of Visual Data Analytics in Erie, Colo., instead plotted data using day of the year as the X axis value, year as the Y axis value, and flow as a Z layer by using the Color Relief mapping option within Surfer. By colorizing flow values, the resulting raster plot clearly showed the magnitude, frequency, duration, timing, and flow change for every day within the entire period of record. [Read our previous blog on this topic here.]
In a follow-up webinar held in February 2020, Koehler demonstrated how to use the same time-series technique with forecast hydrologic data to visualize and quantify possible future flow conditions. He created a raster plot containing dozens of streamflow scenarios, each of which were derived from different future meteorological conditions.
For the recent webinar, Koehler used flow scenarios created for a location on the Colorado River. In this case, the high-to-low ranks of streamflow volumes, rather than simulation year, were used for the Y Axis coordinate. He then displayed all scenarios as a time-series raster visualization in Surfer.
"Among the more interesting results, the visualization showed that in wetter years (higher snowmelt runoff condition), the peak streamflow occurs later, in late June, and higher flows lasts a couple months," said Koehler. "In drier years (lower snowmelt runoff condition), the peak streamflow comes about a month earlier and the higher flows don't last as long."
"These forecasts are valuable to anyone who runs a business that depends on the river" said Koehler. "That would be farmers, water districts, fishing guides and rafting companies."
The predictions give them a chance to prepare for what will happen on the river in terms of water availability in the coming spring and summer, he explained. Farmers can plan accordingly, and rafting companies can decide when to hire their guides and begin their spring trips.
As is always the case with visualization in Surfer, "communicating data visually is useful because people can see patterns for themselves and start asking questions. They don't need a bunch of statistics," Koehler concluded.
Koehler pointed out that time-series visualizations aren't limited to water flow rates. This technique can be applied to most any type of time series data such as salmon migration for fisheries, snowpack conditions for water supply, traffic patterns for urban planning, disease spread for pandemic modeling, marketing effectiveness for business decisions, physiological data displays for medical readouts, and various sports data visualizations.
Watch the entire recorded webinar below: