Catchment-scale estimates of flow path partitioning and water storage based on transit time and runoff modelling

C. Soulsby, K. Piegat, J. Seibert, Doerthe Tetzlaff

Research output: Contribution to journalArticle

39 Citations (Scopus)

Abstract

Tracer-derived mean transit times (MTT) and rainfallrunoff modelling were used to explore stream flow generation in 14 Scottish catchments. Both approaches conceptualise the partitioning, storage, and release of water at the catchment scale. The study catchments were predominantly upland and ranged from 0.5 to 1800?km2. Lumped convolution integral models using tracer inputoutput relationships generally provided well-constrained MTT estimates using a gamma function as the transit time distribution. These ranged from 60?days to >10?years and are mainly controlled by catchment soil cover and drainage density. The HBV model was calibrated using upper and lower storage layers to conceptualise rapidly responding near-surface flow paths and slower groundwater contributions to runoff. Calibrated parameters that regulate groundwater recharge and partitioning between the two storages were reasonably well-identified and correlations with MTTs. The most clearly identified parameters and those with the strongest correlations with MTT and landscape controls (particularly soil cover) were the recession coefficients which control the release of water from the upper and lower storage layers. There was also strong correlation between the dynamic storage estimated by HBV and the total catchment storage inferred by tracer damping, although the latter was usually two orders of magnitude greater. This is explained by the different storages estimated: while the total storage inferred by tracers also includes the passive storage involved in mixing, the model estimates dynamic storage from water balance considerations. The former can be interpreted as relating to total porosity, whereas the latter rather corresponds to the drainable porosity. As MTTs for Scottish the uplands can be estimated from catchment characteristics, landscape analysis can be used to constrain sensitive model parameters when modelling in ungauged basins. Furthermore, the dynamic storage inferred by HBV may also be used to provide a first approximation of minimum total catchment storage. Copyright (c) 2011 John Wiley & Sons, Ltd.

Original languageEnglish
Pages (from-to)3960-3976
Number of pages17
JournalHydrological Processes
Volume25
Issue number25
Early online date15 Nov 2011
DOIs
Publication statusPublished - 15 Dec 2011

Keywords

  • tracers
  • transit times
  • rainfall-runoff modelling
  • runoff processes
  • storage
  • HBV
  • ungauged basins
  • residence times
  • process conceptualization
  • spatial interpolation
  • mesoscale catchment
  • upland catchments
  • stream water
  • rainfall
  • tracer
  • uncertainty
  • topography

Cite this

Catchment-scale estimates of flow path partitioning and water storage based on transit time and runoff modelling. / Soulsby, C.; Piegat, K.; Seibert, J.; Tetzlaff, Doerthe.

In: Hydrological Processes, Vol. 25, No. 25, 15.12.2011, p. 3960-3976.

Research output: Contribution to journalArticle

@article{cbb313f935da423e99b8789e249f8e6b,
title = "Catchment-scale estimates of flow path partitioning and water storage based on transit time and runoff modelling",
abstract = "Tracer-derived mean transit times (MTT) and rainfallrunoff modelling were used to explore stream flow generation in 14 Scottish catchments. Both approaches conceptualise the partitioning, storage, and release of water at the catchment scale. The study catchments were predominantly upland and ranged from 0.5 to 1800?km2. Lumped convolution integral models using tracer inputoutput relationships generally provided well-constrained MTT estimates using a gamma function as the transit time distribution. These ranged from 60?days to >10?years and are mainly controlled by catchment soil cover and drainage density. The HBV model was calibrated using upper and lower storage layers to conceptualise rapidly responding near-surface flow paths and slower groundwater contributions to runoff. Calibrated parameters that regulate groundwater recharge and partitioning between the two storages were reasonably well-identified and correlations with MTTs. The most clearly identified parameters and those with the strongest correlations with MTT and landscape controls (particularly soil cover) were the recession coefficients which control the release of water from the upper and lower storage layers. There was also strong correlation between the dynamic storage estimated by HBV and the total catchment storage inferred by tracer damping, although the latter was usually two orders of magnitude greater. This is explained by the different storages estimated: while the total storage inferred by tracers also includes the passive storage involved in mixing, the model estimates dynamic storage from water balance considerations. The former can be interpreted as relating to total porosity, whereas the latter rather corresponds to the drainable porosity. As MTTs for Scottish the uplands can be estimated from catchment characteristics, landscape analysis can be used to constrain sensitive model parameters when modelling in ungauged basins. Furthermore, the dynamic storage inferred by HBV may also be used to provide a first approximation of minimum total catchment storage. Copyright (c) 2011 John Wiley & Sons, Ltd.",
keywords = "tracers, transit times, rainfall-runoff modelling, runoff processes, storage, HBV, ungauged basins, residence times, process conceptualization, spatial interpolation, mesoscale catchment, upland catchments, stream water, rainfall, tracer, uncertainty, topography",
author = "C. Soulsby and K. Piegat and J. Seibert and Doerthe Tetzlaff",
year = "2011",
month = "12",
day = "15",
doi = "10.1002/hyp.8324",
language = "English",
volume = "25",
pages = "3960--3976",
journal = "Hydrological Processes",
issn = "0885-6087",
publisher = "Wiley-Blackwell",
number = "25",

}

TY - JOUR

T1 - Catchment-scale estimates of flow path partitioning and water storage based on transit time and runoff modelling

AU - Soulsby, C.

AU - Piegat, K.

AU - Seibert, J.

AU - Tetzlaff, Doerthe

PY - 2011/12/15

Y1 - 2011/12/15

N2 - Tracer-derived mean transit times (MTT) and rainfallrunoff modelling were used to explore stream flow generation in 14 Scottish catchments. Both approaches conceptualise the partitioning, storage, and release of water at the catchment scale. The study catchments were predominantly upland and ranged from 0.5 to 1800?km2. Lumped convolution integral models using tracer inputoutput relationships generally provided well-constrained MTT estimates using a gamma function as the transit time distribution. These ranged from 60?days to >10?years and are mainly controlled by catchment soil cover and drainage density. The HBV model was calibrated using upper and lower storage layers to conceptualise rapidly responding near-surface flow paths and slower groundwater contributions to runoff. Calibrated parameters that regulate groundwater recharge and partitioning between the two storages were reasonably well-identified and correlations with MTTs. The most clearly identified parameters and those with the strongest correlations with MTT and landscape controls (particularly soil cover) were the recession coefficients which control the release of water from the upper and lower storage layers. There was also strong correlation between the dynamic storage estimated by HBV and the total catchment storage inferred by tracer damping, although the latter was usually two orders of magnitude greater. This is explained by the different storages estimated: while the total storage inferred by tracers also includes the passive storage involved in mixing, the model estimates dynamic storage from water balance considerations. The former can be interpreted as relating to total porosity, whereas the latter rather corresponds to the drainable porosity. As MTTs for Scottish the uplands can be estimated from catchment characteristics, landscape analysis can be used to constrain sensitive model parameters when modelling in ungauged basins. Furthermore, the dynamic storage inferred by HBV may also be used to provide a first approximation of minimum total catchment storage. Copyright (c) 2011 John Wiley & Sons, Ltd.

AB - Tracer-derived mean transit times (MTT) and rainfallrunoff modelling were used to explore stream flow generation in 14 Scottish catchments. Both approaches conceptualise the partitioning, storage, and release of water at the catchment scale. The study catchments were predominantly upland and ranged from 0.5 to 1800?km2. Lumped convolution integral models using tracer inputoutput relationships generally provided well-constrained MTT estimates using a gamma function as the transit time distribution. These ranged from 60?days to >10?years and are mainly controlled by catchment soil cover and drainage density. The HBV model was calibrated using upper and lower storage layers to conceptualise rapidly responding near-surface flow paths and slower groundwater contributions to runoff. Calibrated parameters that regulate groundwater recharge and partitioning between the two storages were reasonably well-identified and correlations with MTTs. The most clearly identified parameters and those with the strongest correlations with MTT and landscape controls (particularly soil cover) were the recession coefficients which control the release of water from the upper and lower storage layers. There was also strong correlation between the dynamic storage estimated by HBV and the total catchment storage inferred by tracer damping, although the latter was usually two orders of magnitude greater. This is explained by the different storages estimated: while the total storage inferred by tracers also includes the passive storage involved in mixing, the model estimates dynamic storage from water balance considerations. The former can be interpreted as relating to total porosity, whereas the latter rather corresponds to the drainable porosity. As MTTs for Scottish the uplands can be estimated from catchment characteristics, landscape analysis can be used to constrain sensitive model parameters when modelling in ungauged basins. Furthermore, the dynamic storage inferred by HBV may also be used to provide a first approximation of minimum total catchment storage. Copyright (c) 2011 John Wiley & Sons, Ltd.

KW - tracers

KW - transit times

KW - rainfall-runoff modelling

KW - runoff processes

KW - storage

KW - HBV

KW - ungauged basins

KW - residence times

KW - process conceptualization

KW - spatial interpolation

KW - mesoscale catchment

KW - upland catchments

KW - stream water

KW - rainfall

KW - tracer

KW - uncertainty

KW - topography

U2 - 10.1002/hyp.8324

DO - 10.1002/hyp.8324

M3 - Article

VL - 25

SP - 3960

EP - 3976

JO - Hydrological Processes

JF - Hydrological Processes

SN - 0885-6087

IS - 25

ER -