|
" A Forecasting Framework for Complex Systems: "
Palash, Khan Md Wahid
Islam, Shafiqul
| Document Type
|
:
|
Latin Dissertation
|
| Language of Document
|
:
|
English
|
| Record Number
|
:
|
1105525
|
| Doc. No
|
:
|
TLpq2312585929
|
| Main Entry
|
:
|
Islam, Shafiqul
|
|
|
:
|
Palash, Khan Md Wahid
|
| Title & Author
|
:
|
A Forecasting Framework for Complex Systems:\ Palash, Khan Md WahidIslam, Shafiqul
|
| College
|
:
|
Tufts University
|
| Date
|
:
|
2019
|
| student score
|
:
|
2019
|
| Degree
|
:
|
Ph.D.
|
| Page No
|
:
|
314
|
| Abstract
|
:
|
A short (3 to 5-day) to mid-range (6 to 10-day) flood forecasting lead time is desired to increase the flood response and preparedness in flood-prone regions across the world. The major limitation of providing accurate flood forecasts of such lead times is mainly associated with system complexity (e.g., nonlinear interactions with feedback among system elements) and modeling complexity (e.g., uncertainty in precipitation measurement, scale mismatch between model equations and heterogeneity of variables and processes). To generate a perfect (or nearly perfect) model of such a system, one has to model everything (or nearly everything). Yet, as Lorenz (1963) aptly pointed out, in a nonlinear system with feedback, approximate accurate representation of the present does not guarantee accurate forecast of the future due to sensitivity to the initial conditions. How do we, then, understand, model, and manage such a complex system? The main objective of this doctoral research is, therefore, to explore and examine the utility of framing flood forecasting as a complex system problem. To achieve this objective, a data-driven simple flood forecasting scheme using the notion of requisite simplicity with a contextual and mechanistic understanding of river basin hydrology and regional hydrometeorological conditions is proposed. Applying requisite simplicity in this context involves identifying key variables and processes of rainfall-runoff, flow travel time, flood propagation along the river, and developing a mechanism to track their evolution, interconnections, feedback and system behavior. The developed data-based model—for its simplicity in model structure, utilization of easily available data, and more importantly, ease of making it operational for real-time applications—is named ReqSim (Requisitely Simple) flood forecasting model. The ReqSim model shows great promise to provide short (3 to 5-day) to mid-range (6 to 10-day) deterministic flood forecasts for medium to large river basins around the world. It also shows the potential of providing accurate forecasts for local scales by using outputs of the regional or global scale models. Findings suggest that models with requisite simplicity—relying on flow persistence, aggregated upstream rainfall, and travel time—can provide reliable flood forecasts comparable to relatively more complicated methods for up to 10-day lead time. Ease of operationalization and reliable forecasting accuracy of the proposed framework is of particular importance for large rivers, where access to upstream gauge-measured rainfall and flow data are limited, and detailed modeling approaches are operationally prohibitive and functionally ineffective.
|
| Subject
|
:
|
Civil engineering
|
|
|
:
|
Hydrologic sciences
|
|
|
:
|
Water resources management
|
| |