Modeled Insect Richness

To better characterize conditions across flow permanence gradients, NAMC built a modeled insect richness index in 2023 using an interagency network of reference sites with watersheds <7,000 km2. This index is calculated by taking the observed number of insect taxa and dividing that by the predicted number of insect taxa from random forest models. This index uses many of the same reference sites included in the westwide index but additionally includes many additional sites in Arizona. Index performance declines as flow permanence declines so benchmarks are recommended based on the percent of the National Hydrography Dataset stream km within a 25 km buffer of each point that are classified as nonperennial. To apply this index samples should be identified to NAMC’s standard taxonomic resolution, but midges are excluded from calculations of number of insect taxa. Samples should be collected between March 8th and December 19th.

Index # Reference Sites Index Mean Index SD Moderate departure from reference benchmark Major departure from reference benchmark Narrative Conditions
<60 %NP25km 144 1.01 0.20 <=0.92 <=0.76 Good, Fair, Poor
60-80 %NP25km 84 0.99 0.25 <=0.86 <=0.67 Good, Fair, Poor
>80 %NP25km 154 0.99 0.31 <=0.82 <=0.62 Good, Fair, Poor

South western U.S. outline consisting of California, Oregon, Washington, Idaho, Nevada, Arizona, Utah, Montana, Wyoming, Colorado, New Mexico, North Dakota and South Dakota with black dots representing reference sites
Modeled insect richness reference sites

Box plots of catch elevation(m) on a scale of 0 to 3500 with a mean of 2000, mean annual precipitation (mm) on a scale of 0 to 3000 with a mean of 750, temperature (C) on a scale of -1 to 20 with a mean of 6 and log watershed area (km2) on a scale of 0 to 4 with a mean of 1.5
Environmental gradients of reference sites used for the modeled insect richness index and used for NAMC’s model applicability test. Note that watersheds >7,000 km2 were excluded from this index.

  • Method for accounting for natural environmental gradients: Random forest modeling predicting total insect richness
  • Predictors of reference assemblages or metrics: 
    • Number of perennial stream km (NHD Plus Fcode 43006) in a 25 km buffer around the point divided by the number of stream km in that 25 km buffer with NHD Plus Fcode 46006)
    • PRISM climate data - 30-year normal maximum temperature (°C): Annual period: 1981-2010 within the watershed
    • Latitude in decimal degrees at the point
    • Mean elevation of the catchment
    • Day of year
    • Mean of predicted mean annual stream temperature (Jan-Dec) for years 2008, 2009, 2013, & 2014
    • Watershed area (square km) at NHDPlus stream segment outlet, i.e., at the most downstream location of the vector line segment
    • Mean % of lithological nitrogen (N) content in surface or near surface geology within catchment
    • Mean seasonal water table depth (cm) of soils (STATSGO) within watershed
  • % of watershed area classified as ice/snow land cover (NLCD 2011 class 12)
  • Fixed count: 300
  • Minimum count: unspecified
  • Operational Taxonomic Unit (OTU) applied: Yes
  • Rare taxa <0.5 probability of capture included: Yes
  • Documentation: Courtwright, J. L., and C. P. Hawkins. 2023. Assessing stream biological integrity across flow-permanence gradients. Freshwater Science (in review).