Text

Evaluation of parameter uncertainty for two conceptual models in Jiangwan catchment, China

---

The necessity of hydrological information and flow forecasting provision is increasing nowadays due to its various utilizations in wide substances. Considering various data availability, analyzed catchment characteristics et cetera, numerous concept and structure of water balance model are developed. In small catchment, hydrological analysis by using simple lumped model has been long applied and found appropriate to be implemented; while majority of developed countries has owned the capability to use distributed model that consider spatial distribution. In regard to such situation, this study is aimed to evaluate the significance of model complexity – related to model construction, spatial concept and structure – in simulating flow in small catchment. The analyzed catchment is Jiangwan catchment, located in Zhejiang Province, People‟s Republic of China, with available hydrological data (daily rainfall and flow) from 1971 to 1986.
Starting from both HBV and Xinanjiang model parameter calibration, this study is then developed by evaluating the output of each model simulation, before further comparing both model deliverance considering the significance of model complexity. Facilitating the complex simulations, HBV model made use Microsoft excel spreadsheet, while Xinanjiang model is built on Visual Basic, with the connection to Digital Elevation Model (DEM) data and hydrological database. Comparing both model structures, Xinanjiang model is capable to take non-uniform spatial distribution into its account as its main feature, while HBV model used in this study is spatially lumped model.
Both models are found to deliver different parameter characteristics. While HBV model parameter alter every year under certain pattern, Xinanjiang model parameters tend to be more stable. Deep examination of the model parameter uncertainty is explained by correlating it to the actual catchment physical condition. Nash-Sutcliffe that evaluates simulated average flow and relative error which evaluates time-to-time simulation are being used as the objective function, besides further examination of high flow and low flow approach. The Nash-Sutcliffe value delivered by optimized HBV and Xinanjiang model respectively are found at 0.5501 and 0.7300. Xinanjiang model delivers much better simulation based on the relative error with average value of 0.0479, compared to HBV model deliverance of 0.1355. Based on high flow approach perspective, it is found that Xinanjiang model also simulates better with average relative deviation of 5.5%, compared to HBV model that delivers 20.97%. However, both models are found to fail to do low flow simulation, expressed by disagreement between the observed and simulated duration curve, presumably due to the ephemeral characteristic of the river. However, taking the model accuracy for flow less than 70% probability occurrence results 55.36% and 13.52% deviation for respectively HBV and Xinanjiang model. Most importantly, both model parameters are found to be able to describe the actual physical condition of the catchment. Conclusively, both HBV and Xinanjiang model delivers good flow simulation result. Yet, Xinanjiang is selected as the better model.
Keywords: water balance model, Jiangwan catchment, HBV model, Xinanjiang model, model complexity, parameter uncertainty

---

Ketersediaan

Barcode		Tipe Koleksi	Nomor Panggil	Lokasi	Status
tes1625	T/DIG - PMTS	Tesis	TES-PMTS RUS e/15	Gdg9-Lt3 (TESIS DOSEN)	Tersedia namun tidak untuk dipinjamkan - No Loan

Informasi Detail

Judul Seri

No. Panggil

TES-PMTS RUS e/15

Penerbit

Bandung : Joint Master Double Degree Program Parahyangan Catholic University Hohai University., 2015

Deskripsi Fisik

vi, 103 p. : il ; 30 cm

Bahasa

English

ISBN/ISSN

-

Klasifikasi

-

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

WATER BALANCE MODEL

Info Detail Spesifik

-

Pernyataan Tanggungjawab

Steven Reinaldo Rusli

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar