Integration of a Three-Dimensional Process-Based Hydrological Model into the Object Modeling System

Giuseppe Formetta,Timothy Green,Olaf David,Giovanna Capparelli,Riccardo Rigon

doi:10.3390/w8010012

Abstract

The integration of a spatial process model into an environmental modeling framework can enhance the model’s capabilities. This paper describes a general methodology for integrating environmental models into the Object Modeling System (OMS) regardless of the model’s complexity, the programming language, and the operating system used. We present the integration of the GEOtop model into the OMS version 3.0 and illustrate its application in a small watershed. OMS is an environmental modeling framework that facilitates model development, calibration, evaluation, and maintenance. It provides innovative techniques in software design such as multithreading, implicit parallelism, calibration and sensitivity analysis algorithms, and cloud-services. GEOtop is a physically based, spatially distributed rainfall-runoff model that performs three-dimensional finite volume calculations of water and energy budgets. Executing GEOtop as an OMS model component allows it to: (1) interact directly with the open-source geographical information system (GIS) uDig-JGrass to access geo-processing, visualization, and other modeling components; and (2) use OMS components for automatic calibration, sensitivity analysis, or meteorological data interpolation. A case study of the model in a semi-arid agricultural catchment is presented for illustration and proof-of-concept. Simulated soil water content and soil temperature results are compared with measured data, and model performance is evaluated using goodness-of-fit indices. This study serves as a template for future integration of process models into OMS.

Highlights

Based on the work of [1] and the development of the SHE model (e.g., [2]) it became obvious that complex process models accounting for hydrologic fluxes and interactions at small scales produce prognostic information not available in lumped models. [3] and [4] provided accurate descriptions and intercomparisons between coupled surface and subsurface flow models
The simulation period (September–August) was selected in order to show the change in saturation from around 0.22 (m3 ̈ m3 ) to around 0.40 (m3 ̈ m3 )
This was mainly due to intense early-season rainfall wetting fallow soils that were relatively dry over the winter due to crop water uptake in the previous year

Summary

Introduction

Based on the work of [1] and the development of the SHE model (e.g., [2]) it became obvious that complex process models accounting for hydrologic fluxes and interactions at small scales produce prognostic information not available in lumped models. [3] and [4] provided accurate descriptions and intercomparisons between coupled surface and subsurface flow models. Water 2016, 8, 12 energy budgets that other models (e.g., [9,10]) obtain only by external coupling with land-surface schemes. For this reason, calibration of the water budget affects the energy fluxes with feedback to water fluxes, which makes the model calibration process challenging. Calibration of the water budget affects the energy fluxes with feedback to water fluxes, which makes the model calibration process challenging Another constraint of many of the above models is their monolithic structure that limits one’s ability to study, improve, and test them collaboratively. As the GEOtop history shows (e.g., [8,11])

Methods

Results

Conclusion