Data Analysis

Statistical Methods

Fixed Stations

Trends in condition at fixed stations are identified using the nonparametric technique of van Belle and Hughes (1984).  This procedure is based on the Mann-Kendall statistic and consists of a sign test comparing each value (the B-IBI score) with all values measured in subsequent periods.  The ratio of the Mann-Kendall statistic to its variance provides a normal deviate that is tested for significance.  Alpha is set to 0.1 for these tests because of the low power for trend detection for biological data.  An estimate of the magnitude of each significant trend is obtained using Sen's (1968) procedure which is closely related to the Mann-Kendall test.  Sen's procedure identifies the median slope among all slopes between each value and all values measured in subsequent periods. 

Baywide Estimations

To estimate the amount of area in the entire Bay that fails to meet the Chesapeake Bay Benthic Restoration Goals (p), we define for every site i in stratum h a variable y_hi that has a value of 1 if the benthic community fails the goals, and 0 otherwise.  For each stratum, the estimated proportion of area failing to meet the goals, p_h, and its variance is calculated as the mean of the y_hi's and their variance.  Estimates for strata are combined in a weighted average to achieve a statewide estimate.  For combined strata, the 95% confidence intervals are estimated as the proportion plus or minus twice the standard error.  For individual strata, (e.g., each of the 10 strata in 2001) the exact confidence interval is determined from tables.

The document Methods for Calculating The Chesapeake Bay Benthic Index of Biotic Integrity contains the information necessary to calculate the Chesapeake Bay B-IBI, including the latest updates.  If Acrobat Reader is set up on your machine, this document will open when you click on the "pdf" image in the box above.  To save the file to your hard disk (size = 915,253 bytes), right-click (Windows) or option-click (Macintosh) on the link.

B-IBI Calculation Method

The Chesapeake Bay benthic index of biotic integrity (B-IBI) was developed to assess benthic community health and environmental quality in Chesapeake Bay.  The B-IBI evaluates the ecological condition of a sample by comparing values of key benthic community attributes (“metrics”) to reference values expected under non-degraded conditions in similar habitat types.  It is therefore a measure of deviation from reference conditions.

The development of the Chesapeake Bay B-IBI has been described in Weisberg et al. (1997).  In addition, a series of statistical and simulation studies were conducted to evaluate and optimize the B-IBI (Alden et al. 2002).  The results of Alden et al. (2002) indicated that the B-IBI is sensitive, stable, robust, and statistically sound.  New sets of metric and threshold combinations for the tidal freshwater and oligohaline habitats were also developed in Alden et al. (2002) with a larger dataset than was available to Weisberg et al. (1997) for these two habitats. 

The Chesapeake Bay B-IBI is calculated by scoring each of several attributes of benthic community structure and function (abundance, biomass, Shannon diversity, etc.) according to thresholds established from reference data distributions.  The scores (on a 1 to 5 scale) are then averaged across attributes to calculate and index value.  Samples with index values of 3.0 or more are considered to have good benthic condition indicative of good habitat quality.