Ohio Nowcast - What Are The Variables Used To Make Predictions On The Nowcast in 2009?

What are the variables used to make predictions on the nowcast in 2013? Edgewater Beach.

The mathematical models used to make predictions on the nowcast are based on measurements of environmental and water-quality variables known to influence E. coli levels. These are called "explanatory variables" and include measurements of rainfall, wave height, and water clarity.

During 2013, there are three subseason models for Huntington and one backup model if radar data are not available. The subseason 3 model for Huntington was developed with radar rainfall, obtained from the National Weather Service, through a series of data transfers. In the event the radar data are not transmitted successfully, a model was developed as a backup using rainfall data obtained from the National Weather Service site at Hopkins International Airport. The models for Edgewater and Villa Angela were developed using variables generated from a rapid analytical method for E. coli—quantitative polymerase chain reaction (qPCR). If qPCR results are not available, a backup model without qPCR will be used. For Maumee Bay State Park, two models were developed; which model is used is dependent on the direction of flow in Berger Ditch.

At Huntington, the variables used in the models for the 2013 nowcast are as follows:

Subseason 1 (May – June 29): (1) natural log of turbidity, (2) day of the year, (3) lake level change in the previous 24 hours (Cleveland station), (4) wind speed at 8 a.m. (Hopkins Airport), and (5) rainfall in the previous 48 hours, with the most recent rainfall receiving the most weight (Hopkins Airport). The threshold probability for issuing an advisory using this model is 30 percent.

Subseason 2 (June 30 – Aug 9): (1) log10 of turbidity, (2) day of the year, (3) lake level change in the previous 24 hours (Cleveland station), (4) airport rainfall from the previous 24 hours, (4) instantaneous wind code x wind speed (Hopkins Airport). Wind code is 0.1 for no wind, 1 for winds 91 to 269° (SE, S, SW), and 2 for winds 270 to 90° (W, NW, N, NE, E) at the time of sampling. The threshold probability for issuing an advisory using this model is 31 percent.

Subseason 3 (Aug 10 - Sept): (1) the sum of log10 of turbidity and the wave height, (2) lake level change in the previous 24 hours (Cleveland station), (3) the sum of radar rainfall from six 4-km grids surrounding the beach and airport rainfall, both during the previous 24 hours, (4) airport rainfall 2 days ago, and (5) instantaneous wind code x wind speed. Wind code is 0.1 for no wind, 1 for winds 91 to 269° (SE, S, SW), and 2 for winds 270 to 90° (W, NW, N, NE, E) at the time of sampling. The threshold probability for issuing an advisory using this model is 40 percent.

Subseason 3 Backup model without radar rainfall: (1) the sum of log10 of turbidity and the wave height, (2) lake level change in the previous 24 hours (Cleveland station), (3) airport rainfall in the previous 48 hours with the most recent rainfall receiving the most weight, and (4) instantaneous wind code x wind speed (as described above). The threshold probability for issuing an advisory using this model is 40 percent.

At Edgewater, the variables used in the models for the 2013 nowcast are as follows:

All season model that includes a variable generated from the quantitative polymerase chain reaction (qPCR) assay for E. coli: (1) log turbidity, (2) water temperature, (3) airport rainfall (Hopkins Airport) during the previous 72 hours with the most recent rainfall receiving the most weight, and (4) the cell equivalent per 100 mL (CE/100 mL) based on the ∆Ct calculation for the qPCR assay. If the qPCR reaction is inhibited or qPCR was not done, the backup model will be used (see below). The threshold probability for issuing an advisory using the qPCR model is 32 percent.

Backup model without a qPCR variable: (1) the sum of log10 of turbidity and the wave height, (2) day of the year, (3) pH, (4) water temperature in degrees Celsius, and (5) sum of airport rainfall and radar rainfall in the previous 48 hours with the most recent rainfall receiving the most weight. The threshold probability for issuing an advisory using the qPCR model is 32 percent.

At Maumee Bay State Park, the variables used in the models for the 2013 nowcast are as follows:

All season, positive flow in Berger Ditch: rainfall 2 days ago at Toledo Metcalf Airport, the wind code summed for the previous 4 hours, and the turbidity. The threshold probability for issuing an advisory using this model is 25 percent.

All season, negative flow in Berger Ditch: the wind code at time of sampling, turbidity, and the instantaneous discharge in Berger Ditch. The threshold probability for issuing an advisory using this model is 30 percent. For all models, wind code (2 categories) is based on wind direction from Toledo Metcalf Airport—N to E winds are coded as “1” and all other wind directions and no wind are coded as “0.1”.

For the Cuyahoga River at Jaite, a model based on turbidity as measured by the in situ turbidity sensor and rainfall as observed at Akron-Fulton International Airport for the day prior to sampling will be used for the nowcast in 2013. On days when the predicted E. coli concentrations in the river are below 298 colony-forming units per 100 milliliters, the water quality will be considered “good”.

For the Cuyahoga River at Jaite, a model based on measured turbidities will be used for the nowcast in 2012.

At Villa Angela, the variables used in the model for the 2013 nowcast are as follows: All season model that includes a variable generated from the quantitative polymerase chain reaction (qPCR) assay for E. coli: (1) wave height, (2) water pH, (3) air temperature, and (4) the calibrator cell equivalents per 100 mL (CCE/100 mL) based on the ∆∆Ct calculation for the qPCR assay. If the qPCR reaction is inhibited or qPCR was not done, the backup model will be used (see below). The threshold probability for issuing an advisory using the qPCR model is 35 percent.

Backup model without a qPCR variable: (1) turbidity, (2) day of the year, (3) air temperature, (4) wave height, and (5) Burke Lakefront Airport rainfall from the previous 48 hours with the most recent rainfall receiving the most weight. The threshold probability for issuing an advisory using the qPCR model is 35 percent.

Ohio Nowcast model variables.