Why Understanding the Hydrology of Rivers is Important

Rivers have been an essential part of human development historically, as they provide a water supply and transportation that have advanced civilization. However, it is because of the settling in close proximity to rivers that they have been greatly affected by human activity (Berner and Berner 2012). The quality of a river is inextricably linked to the health of the population settled in close proximity to it, and so it is imperative to learn about the hydrologic fluxes in a river and how they relate to the water quality.

This research project focuses on 6 rivers that fall within the watersheds of the Upper Hudson and Mohawk Rivers. The rivers being studied are the Upper Hudson, Wampecack Creek (WC), the Mohawk, Nowadaga Creek (NC), Otsquago Creek (OC), and Caroga Creek (CC). The aim of this project is to take a comprehensive approach to characterizing these rivers so we can better understand how the water quality is influenced. There is large variability in size and lengths of the rivers and watersheds which allows for us to further investigate how factors such as hydrology, geology and land use of the watersheds affects the geochemistry in the rivers and reservoirs that feed them. By better understanding the general trends of the region of our rivers and how they compare to the rest of New York, the rest of the country, and the rest of the world, we are able to frame our results and place their meaning into a better understood context. In conducting hydrograph analysis, hydrologic flux calculations, and geochemical analysis we are able to understand the controls on river water quality and highlight possible anthropogenic inputs that are not explained by watersheds characteristics. Comprehending the composition of a watershed and the chemistry of the water within its boundaries is crucial to ensuring the health of the watershed and all of the organisms that use and inhabit it. This project hopes to provide a model for the characterization of a watershed and stream chemistry that can be translated and expanded to larger regional hydrology.

Geology and Land Use practice

The land use practices vary greatly from watershed to watershed (Table 1). Land use data was compiled from NLCD 2011. The land use practices in a watershed is linked closely with the geology f a watershed. For instance across the studied rivers the majority of the agriculture practices are concentrated in watersheds that are underlain by sedimentary formations and OM rich shales like those of WC, OC, and NC. (Longworth et al. 2007). The land use practices and geology of a watershed have a strong control over the geochemistry in a river and this will be discussed later. The other rivers that are comprised of mainly forested land have lithologies of either crystalline rock (CC), or the mixed geologies of the Hudson and Mohawk which combine crystalline rocks from the Adirondacks with pockets of other geologic formations from its sub watersheds. The impact that land use practices and geology have on the geochemistry will be discussed later.

Table 1: Land Use Percentages for Each Watershed

Hydrology

Data collection and Analytical Methods

In order to characterize the hydrology of each watershed I compiled large data sets from multiple organizations and agencies. Some of the datasets include precipitation gathered from NOAA and the EPA, streamflow data from USGS stream gauges, runoff data collected from the Waterwatch by the USGS, and land usage data from National Land Cover Database 2011. These datasets were also cleaned and organized to be loaded into R. It is in R that I construct hydrographs, study the different hydrologic fluxes, and characterize the watersheds through the data analysis and data visualization was conducted. The data analysis included calculating runoff ratios, statistical analysis on monthly and annual streamflow, and calculations for the different fluxes, flow exceedance curves, and shifts in flow timing.

Establishing the General Climate

Climatic variables like precipitation and evapotranspiration have immense control over the hydrology of a watershed. So, I looked at the precipitation levels in the United States and New York State to put our watersheds into context. U.S. precipitation vary greatly across the United States, but the nation average is 76.5 cm annually. New York state ranks 25th out of the states in the U.S. in precipitation receiving an average of 103.8 cm of precipitation annually. This is consistent with the regional precipitation of the Northeast (NOAA). This could be shifting as the average precipitation has been increasing in both New York and Nationally (NCA). This can be seen in Figure 1 and Figure 2

Figure 1

Figure 2

However, it is the combination of precipitation and E.T. which result in the impact we see in stream flow. When E.T. is very high like in the warmer months when primary productivity increases, little precipitation will become result in streamflow becoming runoff. In Figure 3 and Figure 4 we see the monthly relationship between the precipitation and E.T. on the Hudson and Mohawk watersheds.

Figure 3

Figure 4

It is the combination and relationship between these two climatic fluxes and the overall hydrology of a watershed that generate the runoff ratio for a river.

Runoff Ratios and Hydrograph Analysis

Runoff ratios are calculated by taking the runoff of a river and dividing it by the precipitation. The runoff data was collected from USGS and paired with the NOAA data to calculate the runoff ratio for U.S. as whole and New York. The average annual runoff ratios came to be 0.542 for New York state and 0.307 for the Continental United State. In term of an even larger spatial reference, North America has a runoff ratio of 0.38 and the global ratio is 0.46 (Berner and Berner 2012). I then calculated the runoff ratios for each of the watersheds by pairing the EPA precipitation data and the modeled runoff values for each watershed. Figure 5 shows the monthly runoff ratios for each watershed.

Figure 5

The seasonal variability can be with all of the rivers generally follow a similar pattern of increasing ratio going into the winter months and reaching its peak through the spring and drops to very low values in the summer months. These changes show the relationship between the climate and runoff ratios, as the lowest runoff ratios come in the summer months when there is still substantial precipitation.

The seasonal variability can be seen when looking at streamflow on the three Rivers that had streamflow gauges (Hudson, Otsquago, and Mohawk). Figure 6 highlight the 25^th , 50^th , and 75^th flow percentiles and the seasonal variability that precipitation, temperature, and E.T. have on the streamflow.

Figure 6

The flow size of the river and watershed also have an impact on the streamflow. We can see that the Hudson and Mohawk Rivers are much larger and can support much larger flow than Otsquago (Figure 7)

In order to look at broader trends over time I looked at the annual stream flow statistics for the gauge’s locations. The gauges have an extensive record and allow for historic droughts, shifts in flow timing, and shifts in flow loads to be seen

Hudson Hydrograph Analysis

Annual Minimum Streamflow

We can see that the annual minimum flow in the Hudson river is increasing sharply over time.

Annual Mean Streamflow

It can be seen that the annual mean streamflow is also increasing in recent history

Annual Max Streamflow

Interestingly the annual max flow has not changed greatly and even appears to be trending downward in recent years. How can this be if the mean annual flow is increasing.

Annual 7-day low flow

The 7 day low flow is the 7 day period with the lowest flow. It is a good indicator of real increases in minimum flow and helps to explain why the increase in mean flow has go up so much while the max has gone down. These charts however to not tell us how much flow we can expect to get a t a given time. In order to that we need to conduct flow exceedance curves.

Flow Exceedance Curve

Flow Exceedance Curves show the level of flow a stream gets under a certain percentile. In other words, it tells us how much of the time we can expect to get a certain level of streamflow. However, this figure does not display the shifts in annual flow amounts that we discussed earlier. To do this we need to partition the flow exceedance curve into time periods.

Partitioning the Flow Exceedance Curve

Here we can see that the increased minimum and mean flow since the 1960s, while the lower 25th percentile of extreme high flow has not changed very much. However, one again we truly cannot see the magnitude of this shift.

Magnitude of the Flow Exceedance Shift

Here we see that the high flow (low percent exceedance), has not changed really and is at a 1 to 1 ratio roughly, but the low flow (high percent exceedance) has increased at a ratio of upwards of 6.

Mohawk Hydrograph Analysis

Annual Minimum Streamflow

We can see an increasing minimum streamflow after the 1960s. This is consistent with the observations of the extreme draught in the 1960, then followed by a wet decade in the 1970s and has continued to increase since (Darmer 1987).

Annual Mean Streamflow

Once again, we see an increasing mean streamflow from the 1970s onward also consistent with the increase in precipitation seen in New York State in Figure 2

Annual Max Streamflow

The max flow has stayed consistent which indicates that the increase in mean flow is coming from a shift on the minimum flow end as the max is exerting no control over it.

Annual 7 day low flow

The 7 day low flow has been increasing since the 60s, but shoes some very low values around the 2000 which is consistent with a severe drought in 2001 which would cause the 7 day low flow to drop considerably.

Flow Exceedance Curve

Partitioned Flow Exceednce Curve

The partitioning of the flow exceedance once again shows the shifts in middle and high flow percentiles consistent with the increase in minimum and 7 day low flow as well as the mean streamflow.

Otsquago Hydrograph Analysis

Annual Minimum Steamflow

While Otsquago has less observations than Hudson and Mohawk do trends can still be drawn from the hydrographs. For example, we see here the extreme drought that swept through the region in the 1960s talked about above.

Annual Mean Streamflow

Once again, the drought in the 60s is displayed and the mean flow appears to be on an upward trend

Annual Max Streamflow

The max flow like the other gauges streams has stayed consistent or decreases slightly

Annual 7 day Low flow

The drought is observed with the increase in the 70s as precipitation increased and stayed elevated.

Flow Exceedance Curve

Time Shift in Flow Exceedance Curve

Because of the lack of points in OC the partitioning has few points before 1960 and the majority of it points in the 1960s during the drought which skews the data and represents the middle flow percentiles as being higher before 1960 than after which is the opposite from what we have seen in the other hydrograph analysis.

Concentration Discharge Curve Analysis

Now that the hydrograph analysis has been completed and the watershed has been roughly characterized, I then turned to looking the chemistry of the watersheds. A concentrations discharge curve is just as it sounds it is a plot of how the concentration of chemical constituents’ changes based on the discharge of the river. This is important to look at because it can help to indicate what the total load of an ion is over the course of a year. We can see if an ion consistently stays at the same concentration no matter the discharge, or if it is diluted as discharge goes up. This gives us a great deal of explanatory power when talking about the sourcing of an ion. In Figure 8 we can see the concentration discharge curves for all of the chemical constituents we studied across the watersheds of interest.

Figure 8

Here we can the conc. discharge for each constituent for each location. This is nice to compare concentrations across different watersheds as well as compare watershed to watershed. We can see constituents that appear to be chemostatic which means that their concentrations essentially stay the same as discharge increases, and we can see some that display some dilution. This can be better seen when the slopes are fitted to these curves.

Figure 9

With the fitted slopes we are able actually see to what extent there is dilution or chemostatic behavior. The closer it is to 0 the less change in concentration, and the closer to -1 the slight decrease in concentration. It is actually at -1 where we see the concentration decreasing faster than the flow increasing. This graph give a great deal of explaining power about the influence of geology on the geochemistry, as we are able to pair constituents that are linked in lithology and would dissociate with each other when weathered.

Biggest take away from my thesis work

When I started to analyze the water chemistry and in looking at the historic levels of the ions in our rivers compared to the field samples I had collected over the summer during research, the thing that stuck out the most was the extremely high elevated levels of Chloride. I first noticed these high levels during baseflow conditions which means that the stream is being fed by majority groundwater. When Characterizing the watersheds there is no explanations for the levels of Cl that we see other than road salt pollution. Road salting as a way to de-ice roads started in the 1940s a little and since the 1960s it has become a regular practice with increasing severity (Dugan et al. 2017). The fact that these levels are in the groundwater and not in a spring flushing event of quick runoff event where it was swept into the rivers is concerning. Groundwater takes a very long time to replenish and to bring these chloride levels back down to their historic levels would take years or decades if road salting was stopped completely.

The black line and points represent the historic Cl measurements and the grey line and points represent the K historic measurements. The dotted red line shows the chloride levels we measured under baseflow and the blue dotted line show the potassium levels that we measured over the summer. K and the other constituents that we measured stay essentially at their historic levels while only chloride has jumped the way that we have seen do to road salting.

References

Berner, K. A. B., and R. A. C. Berner. 2012. Global Environment: Water, Air, and Geochemical Cycles. Global Environment: Water, Air, and Geochemical Cycles (Second Edition). https://doi.org/10.1590/S0102-695X2013005000004.

Longworth, B. E., S. T. Petsch, P. A. Raymond, and J. E. Bauer. 2007. “Linking Lithology and Land Use to Sources of Dissolved and Particulate Organic Matter in Headwaters of a Temperate, Passive-Margin River System.” Geochimica et Cosmochimica Acta. https://doi.org/10.1016/j.gca.2007.06.056.

https://stahlm.github.io/ENS_215/Data/NOAA_State_Precip_LabData.csv

https://nca2018.globalchange.gov

Darmer, Kenneth I. 1987. “Overview of Hudson River Hydrology.

Dugan, Hilary A., Sarah L. Bartlett, Samantha M. Burke, Jonathan P. Doubek, Flora E. Krivak-Tetley, Nicholas K. Skaff, Jamie C. Summers, et al. 2017. “Salting Our Freshwater Lakes.” Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.1620211114.