View the SRWP Facebook Page  View the SRWP Twitter Feed  Signup for the SRWP Email Newsletter

6.1 Critical data gaps

Indicator selection and analysis is often complicated by gaps in the necessary information or data. These gaps exist because either the information has not been collected, or the information exists but is not accessible for any number of reasons. Being aware of the critical data gaps not only helps readers of the Report Card understand the limitations behind the analysis but also provides guidance on any next steps for additional investigations or monitoring. The following summarizes the key issues faced and recommendations regarding various aspects of data availability.

  1. Spatial uniformity of data coverage: It is difficult to carry out analysis for a study area using data that are only representative of a small portion of it (i.e., if data have not been collected for the entire study area). Data that have only been collected at a single location in a single subwatershed cannot be used to generate an indicator that is intended to represent condition across all subwatersheds. Expanded spatial coverage of data collection for specific indicators, and availability of such data, would have made analysis more comprehensive. Inadequacy of spatial representation across subwatersheds was especially true for:
    • Periphyton. Data periphyton biomass and percent cover do not exist for the lower watersheds, i.e., Lower Yuba, Lower Bear, and Lower Feather. There were also no data for the Upper Bear. Periphyton data that were available were collected by four independent organizations: Friends of Deer Creek, the SYRCL, Sabra Purdy (UC Davis) and Fraser Shilling (UC Davis). The limited data that were available were fairly patchy and only collected at one or two sites within a subwatershed: Deer Creek, Middle Yuba, South Yuba, and East Branch North Fork and the North Fork of the Feather. As a result, no trends or complex statistical analysis could be run for periphyton across the region.
    • Benthic macroinvertabrates. Given the limited hydrological mapping of the study area, it was not possible to identify stream order for all benthic macroinvertebrate sampling sites. Stream order is a particularly important variable when analysing macroinvertebrate data because the size of stream influences the composition and structure of the macroinvertebrate community. Better resolution of hydrological maps capturing stream order (i.e., capturing 4th, 5th, and 6th order streams) would have enabled the benthic macroinvertabrate functional feeding groups to be analyzed within a natural continuum of habitat conditions based on ascending stream order.
  2. Frequency of data collection: Some data were only collected at a single point in time and therefore only represent a snapshot of the biotic condition. It is not possible to infer a trend (i.e., whether things are getting better or worse) from a single sampling event. For example, a monthly measurement of a process such as water flow which changes continuously limits the level of analysis that is possible as well as limiting the confidence in results. More frequent sampling of select indicators, and having these data available to analysts, would have enabled a more informed analysis. This was especially true for:
    • Water temperature. The most spatially-extensive temperature data available for the study area were collected once a month by volunteers. This is problematic because the fluctuations in water temperature (e.g., the hottest and coldest water temperatures on that day) are not captured. Similarly, you would not know the length of time that water temperatures remained very high or very low. Maximum temperatures are significant for the aquatic ecosystem because high water temperatures that last a long time can cause stress for fish and insects, sometime to the point of death. Temperature measurements taken every 15 minutes using a continuous recording instrument, such as a HOBO data logger, would enable a more reliable and accurate estimation of the extent and duration of maximum water temperatures. Temperature data from a data logger were only available for a small number of sites and for only a short time period, 2008-2009. To determine trends in maximum water temperature, a continuous time series of ten years or more is needed. Consequently, our analysis for the study area only provide a limited sense of how water temperatures have changed over time and we were not able to determine the duration of high temperature conditions during biologically critical times of year. Temperature data are inexpensive to collect using automated data loggers and do not require a lot of complex equipment, therefore improved data collection should be encouraged.
    • Economics. The school lunch program enrollment data are currently only available as an annual statistic (i.e., X number of children, per school, per year). Monthly data on enrollment in the program would allow for detection of changes in poverty at finer scale to see if there are any relationships between poverty and seasonal patterns of employment and migration of workers.
  3. Availability of reference standards: Control sites in a pristine watershed would be very useful to determine appropriate standards for watershed condition as measured by an indicator. For example, if a pristine watershed was known to have a certain type of benthic macroinvertebrate community in small streams, we could use this as a benchmark to evaluate the health of benthic macroinvertebrate communities in the study area. A lack of appropriate reference conditions was especially true for:
    • Benthic macroinvertebrates. Information on invertebrate species richness from a ‘pristine’ watershed would have been extremely valuable for determining how to score subwatershed in our study area. The ‘pristine’ watershed would have provided a benchmark condition that all subwatersheds could be ranked against.
    • Flow. Historical flow records for the study area have numerous limitations because of incomplete historical records for some subwatersheds. For example, flow data prior to the building of dams in the Upper Feather River (Butt Valley Reservoir, Beldon Reservoir, Rock Creek Reservoir, Cresta Reservoir, Poe Reservoir) do not exist. Collection of flow data for these systems began at the same time the dam was constructed. As a result, we were not able to compare current flows to the natural hydrograph (i.e., what happened before the dams) for the North Fork of the Feather River. Reconstruction of the naturalized flows for the system (for each of the subwatersheds) would provide a better benchmark to compare existing hydrologic conditions.
  4. Consistency in data collection methods: The way data are collected (i.e., how sampling is done) affects the results; if the same metric is sampled in two different ways it is very difficult to compare the results. This was especially true for data available to us on:
    • Benthic macroinvertebrates. The taxonomic level at which an individual bug is identified, typically either genus or family, and rarely down to species, can determine the level of analysis that is possible. The use of a sampling framework across the region with standardized sampling protocols and a wide distribution of sampling locations would have improved the benthic macroinvertebrate analysis because the creation of certain indices is only possible when species-level information is collected.
    • Fish. The method used to sample fish, e.g., electroshocking or snorkeling, determines the number or species and individuals one is able to record. Only one of the fish surveys was thorough enough to look at total community structure whereas the others typically only reported catching 5-6 species (which is probably an artifact of the methodology rather than a true absence of more species).
    • Periphyton. Periphyton data collected using different field and laboratory methods can result in different results. If periphyton are collected using similar protocols, such as the ones in the CWAM (Volume II), or those published by SWAMP, then trend analysis over time and between different watershed would be more reliable. However, even standard protocols like the ones listed have limitations when very large algal masses are encountered.

A few additional data gaps came to light as analyses were conducted:

  • Field Data: There is a need for channel cross-section data in order to determine the stage discharge relationship associated with flooding at different locations (i.e., how much water needs to be flowing through a given location for the riparian area to become inundated). This information would help identify restoration sites where flooding could be allowed without damaging infrastructure.
  • Community Awareness: Public support of ecosystem protection and restoration, public health, and sustainability is a critical feature of a modern society, but it is hard to measure because of the absence of data and data-collection programs. Support begins with awareness. By measuring awareness of watershed health issues, support can be generated for the life-support functions of watersheds.

Addressing the critical data gaps listed above would greatly strengthen the regional analysis as a whole and should serve as a to-do list for those interested in continuing the Report Card analysis into the future. Of utmost priority are:

  • Water temperature measurements taken using automated data loggers, set to 15 minute intervals, and placed strategically across the watershed.
  • Natural flow data generated using a hydrologic model so that current conditions can be compared to a more reliable and standardized data set.
  • Periphyton data collected using standard methods in the upper and lower watershed for the Yuba, Bear and Feather River Watersheds.
  • Standardized and regular monitoring of fish and wildlife populations, probably the best indicators of watershed conditions.

The use of standardized protocols and creation of regional data sets will be made possible only with regular communication between research organizations, regulatory agencies, and watershed groups, because in the end, the management decisions that are influenced by the above analysis are only as good as the data that are collected and made available to collaborative efforts such as this.