Modules


Data Gathering Module

"a generic and modular software for real-time flood forecasting & warning"

Gathers data from a variety of sources and prepares the data in a form suitable for use by the FloodWorks forecasting module. Data sources include:

  • Telemetry networks
  • Hydrometric databases
  • Weather radar
  • Satellite imagery
  • Meteorological forecasts

Each run of the data-gathering module creates a 'snapshot' of the input data.

For the purpose of testing and training, the data-gathering module can also be switched to gather historical event data while simulating real-time operation.

top

Forecast Module

"a generic and modular software for real-time flood forecasting & warning"


Flexible and efficient forecast

Uses a network of models to calculate forecasts of flow, level and other variables at distinct 'forecast points' throughout a river basin.

Combines many types of model, including hydrological and hydrodynamic, into a single, integrated network.

Combines outputs from each model with observed data to update the model’s forecasts and provide the input to models farther downstream.

Provides forecasts for selected parts of the region only, using a single, integrated model network for the entire region (e.g. produce forecasts for upstream reaches only using fast, hydrological models).

Designed for real-time use, particularly in terms of speed of operation and robustness to missing or invalid data.

Validates and cross-checks input data, and in-fills missing or invalid data by specialised models within the network.

Data manipulation and preparation

Provides a variety of simple model algorithms to manipulate and prepare data for the more complex hydrological models. These include algorithms to:

  • validate and cross-check data from different sources to detect instrument errors
  • calculate areal-average rainfall from point measurements
  • calculate areal-average potential evaporation from point measurements of temperature and humidity
  • combine observed and forecast data from different sources according to user-defined priorities
  • convert between level and flow

State updating and error prediction

Compare observed data with the corresponding model outputs and use this information to update and improve the model forecasts:

State-correction: corrects the states of simple models with a small number of internal state variables (such as rainfall-runoff models) using an empirical scheme to apportion the model error between the states.

Error-prediction: provides auto-regressive moving-average (ARMA) models to predict error for complex models with many states and multiple outputs (flow routing, hydrodynamic models). The predicted errors are then subtracted from the model output to give the updated forecast.

top

Analysis Module

"a generic and modular software for real-time flood forecasting & warning"


Forecast Analysis

Generates time series of flow, level and other data at forecast points throughout the FloodWorks model network.

Automatically summarises and interprets these time series with respect to warning criteria defined for the different forecast points and for the region as a whole.

Enables calculation of inundated areas from forecasts of river levels.

Produces a set of tables containing geo-referenced summary data from the forecast that include:

  • cumulative rainfall for all catchments over user defined periods before the forecast origin
  • cumulative rainfall for all catchments during user defined periods that end after the forecast origin
  • the number of missing and/or invalid data values in the period up to the forecast origin, for each station providing input data
  • the highest forecast flow, level or inflow volume at each forecast point, depending on the forecast requirement during user defined periods after the forecast origin (plus the time at which the level was reached)
  • the highest forecast warning status at each forecast point during user defined periods after the forecast origin (plus the time at which the warning was first exceeded and the forecast duration of the exceedance

Forecast Database

Stores the time-series forecasts generated by the FloodWorks Forecasting Module and the summary tables produced by the Analysis Module, in the FloodWorks Forecast Database located on the main server.

The forecast database holds results from each run of the forecasting system independently, together with a copy of all the input data used by the run.

top

Model Algorithms

"a generic and modular software for real-time flood forecasting & warning"


Snowmelt Model

The standard choice for rainfall-runoff modeling in a FloodWorks model network is the Pragmatic Snowmelt (PACK) model developed by CEH Wallingford:

  • the snowmelt process is represented using a snow store and a melt store
  • melting of the snow store is controlled by a simple temperature index equation
  • the resulting melt enters the melt store where it is released slowly from its base
  • a second higher orifice allows release of water from the pack at a higher rate
  • the height of the orifice varies with the total water equivalent of the pack
  • this serves to represent the rapid break-up of the pack as a critical liquid water content is reached
  • an additional component is included to allow for incomplete spatial coverage of snow over a catchment for shallower, older packs
  • this employs an areal depletion curve to calculate the proportion of the catchment covered by snow, allowing some rain to fall on snow-free ground and effectively enter the rainfall-runoff model directly

Rainfall - Runoff Model

The standard choice for rainfall-runoff modeling in a FloodWorks model network is the Probability Distributed Moisture (PDM) model developed at the CEH Wallingford.

The PDM models within a FloodWorks network carry out continuous accounting of soil moisture both between events and from season to season.

The PDM model is suitable for forecasting over the full range of flows and has been used in operational drought management systems as well as flood forecasting systems.

Flow - Routing Model

The KW model, developed at the CEH Wallingford, is the standard model for flow routing within the FloodWorks, and uses the kinematic wave approximation to the St Venant equations to represent simple channel flow that is not subject to backwater effects.

Lateral inflows at each node of the model may be input independently or scaled from upstream flow, and various simple functions may be applied to the flow at each node to represent the effects of out-of-channel storage or loss.

Wave speed may be varied as a function of flow

There is no state-correction within the model; updating is achieved by applying one or more ARMA error-predictors to the model outputs.

The KW model is able to provide fast and effective modeling of flow for the relatively steep, upstream reaches of a river network.

Hydrodynamic Model

ISIS Flow, the flow module from the ISIS suite of river modeling programs, provides hydrodynamic modeling capabilities for the FloodWorks.

It exists both as a standalone program (for offline model building, calibration and simulation) and as a model algorithm within FloodWorks (for real-time, operational forecasting).

The hydrodynamic calculation code and the model data files are common to these two modes of operation and an extensive range of data manage and model building tools are available through InfoWorks-RS.

top