contents | basic aspects of hydrology | hydrological processes | sample processes | catchment | rain discharge transformation within a catchment | hydrology of floods | what is the problem with hydrology | hydrological modelisation | typical problems for FIRMA model (with example) |

basic aspects of hydrology

the aim of hydrology:

To determine how much water will be in a given location and condition. This is a more complex issue than it at first might seem; for example, as local mayor may want to address local flooding, but this is not a purely local issue, and the areas up and downstream need to be considered: there is the need then to 'zoom out'.

The hydrological cycle: Is a continuum, broken by the observer into

• storages
  • water bodies
  • with possible internal evolutionary laws
• water fluxes
  • inside or between water bodies
  • associated to hydrological processes

The main freshwater storages are: Ranked here by increasing time constant - atmosphere

• soil moisture (non saturated area)
• rivers
• snowpack
• lakes ; reservoirs
• groundwater (saturated area)
• icepack

The main freshwater fluxes include:

• Precipitation
• (actual) Evapotranspiration
• Infiltration and seepage (= ex filtration)
• Runoff (on slopes)
• Discharge (in rivers)

hydrological processes:

• Water fluxes are linked to hydrological processes:
  • not only fluxes between water bodies
  • also internal evolution of water bodies
• A process is an elementary behaviour that can described as a whole, and whose level of formalisation may vary. It is under control of various factors

sample processes

Sample process: runoff formation

• according to Horton, runoff occurs where and when rain rate exceeds infiltration capacity
• according to Capus, Hewlett, Beven, runoff occurs where and when rain falls on saturated areas
• importance of the soil structure

Sample process: runoff collection to discharge

• Overland flow (on slopes)
  • Gullies, connectivity topics
  • Importance of relative location of land use
  • Importance of subrogate features of land use (direction of ploughing)

Sample process: underground flow

• The continuous model
  • unsatured zone : the Richards equation
  • satured zone : the Darcy equation
• local formula
• integrated form for alluvial aquifers
• integrated form for constrained aquifer
  • The problem of parameters estimation
• importance of K(q, x, y, z) (a tensor)
• Preferential pathes
  • biological macropores
  • pipes
  • roots
• Impervious layers
  • bottom of ploughing area

catchment

An key hydrological object is the catchment (= the basin)

• An outlet
• The river network upstream
• Slopes - both side of the rivers
  • up to the water divides
• Includes
  • surface and subsurface storages in relation to the river

• the best possible system to study as far as geophysical fluxes as considered
  • one input (rain, other atmospheric conditions)
  • one output (discharge at the outlet)
• the best possible unit for effective management
  • what I do here is my problem

• A fully explicit, exhaustive description is impossible because of
  • the fractal nature of the river network
  • the fractal nature of the topography
  • the partially unreachable description of the under ground
  • the unsteady character of the topography and soil properties at detailed scale
• The catchment is
  • a point (the outlet)
  • a set of lines (the river network)
  • an area (interacting with the atmosphere),
  • a volume (including the underground). o Implementation of such an object in a GIS is not straightforward.
• The definition of a catchment is outlet dependent.
• Two gauging stations define either nested or non-nested catchments
• Data out of many catchments are part of a data hierarchy that must sometimes be considered explicitly (discharge mapping).
• Some problems seem point oriented...
  • ´ how can I reduced floods here -a Ö but must be handled considering all the processes - upstream (causes) and
  • downstream (consequencies of options to take). Often we have to ´ zoom out -a to have a grasp at the problem as a whole.
• It is usually not an administrative division
• The concept may break down - karstic areas - flat, human dominated areas

hydrological monographies

• A balanced description of a catchment (hydrological monography) can be very interesting.
• It will not solve all possible and unexpected questions. A problem oriented description of a catchment
• needs a variety of choices to be done
  • selecting the processes relevant to the problem
  • the scale of the features to explicitely take into account.
  • the time to be considered
• the abduction of non-relevant details has to remain in mind.

Production and transfer functions

• ´ Production -a
  • relates the gross precipitation over the catchment to the net precipitation that is to flow through the outlet.
  • non-flowing water is only considered as a soil moisture controlling factor, influencing the soil behaviour under further rains.
• ´ Transfer -a
  • relates the produced ´ net precipitation -a to the discharge.

About this scheme

• It is common choice to
  • - upload the production function with all the non-linearity of the rain-discharge transformation.
  • -consider the transfert function as linear.
• This approximation may be valid for heavy rains

conceptual approaches to the transfer aspects

• Unit Hydrograph (Sherman, 1932)
  • the transfer function is assumed linear.
  • the structure of the non-linear production function remains author-dependant.
  • parameters for both parts are identified from a joint pair of long rainfall/discharge time series.
• Geomorphologic Unit Hydrograph :
  • an improvement from the previous approach
  • the shape of the unit hydrograph is related to distances and slopes along the runoff pathways from the catchment to the outlet
  • this gives clearer constraints to what the production function can be

limits to these approaches

• Isotopes evaluations show that most of the water of the flood has been in the soil long before the begin of the rain.

• Continuity equation
• Dynamic equation
  • Head
• potential energy + kinetic energy
  • Head losses o along the stream (energy loss in turbulence, interactions between the water and the reach)
• localised (in hydraulic jumps from torrential to fluvial conditions

various levels of description for hydraulic transfer

• in general, PDE equations
• 3D equations (Navier Stokes)
  • small scale studies like geomorphology, flow around a bridge
• 2D equations (BarrÈ Saint-Venant)
  • where overland flow is most relevant : dam breakes, flooding of broader areas with non negligible speed in the flooded part
• 1D+storages (Barre Saint-Venant) :
  • where the flooded area is broken in independent storages, where speed is negligible
• 1D (Barre Saint-Venant) :
  • where streamflow is concentrated in the minor riverbed (no flooding).
  • including dam breaks, working spillways, moving hydraulic jumps, ...
• Simplified 1D equations :
  • Diffusive wave approximation :
• flood diffusion in gentle, sub-horizontal rivers
  • Cinematic wave approximation :
• flood propagation in steep rivers or lateral slopes
• 1D, steady-state approximation :
  • if time variations are negligible. Mostly broad, gentle rivers,
• a important step for text-books in hydraulics (clear, intuitive relation of results to energetic consideration and limits)
• 1D, uniform approximation :
  • to be considered only in regular, chenalized reaches

• To predict floods, or to assess flood hazard?
  • To predict o Given a current stage of water and observed or predicted rain, guess the shape, time of arrival and water stage to occur in the next future at the interest point.
  • To assess o Given a observed discharge time-serie, give probability of a given flood characteristic (peak flow, duration, volume) to be over-seeded

• who
  • civil servants; river authorities; mayors; meteorologists; hydrologists
• how
  • real time data collection
  • quick data processing, mostly empirical models or analogues
  • 365 days, 24 hour communication system to people
• what
  • technical choice of a flood index to predict, level of confidence
• to who
  • police, municipality representatives, everybody ?
• what to say
  • how clear the warning messages ?
  • readiness to cooperate ?

a personal interpretation

• some rivers have long time constants
  • gentle rain, so progressive saturation ; broad basins, so long hydraulic transfers
• some rivers have short time constants
  • steep, small catchments ; convective storms.
• this
  • enable different kind of human measures
  • induced an "hydrology of flash floods" to exist
• but hydrology is one!

flood management approaches

• flood
  • is a natural event
  • can be characterised as an random event => alea
• flooding
  • can yield damages
  • this depends on the sensitivity of land use vulnerability

the damage approach : principle

• considers vulnerability as the cost of damages
• to minimise by - protective measures (levees),
  • storing or evacuating waters via various works,
• as far as monetary evaluation proves efficient.

the damage approach : drawback

• Due to ...
  • the probabilistic nature of events,
  • the short memory of human beings,
  • teleconnections of local actions and basin-wide effects,
• spontaneous local management exhibits a drift towards heavy works that appears to be unsustainable at the basin scale (spiral of corrective measures).

the alea / vulnerability approach

• vulnerability of each type of landuse is a socially determined, possibly negociated acceptance for flooding
• some areas, like marshes, may have a positive demand for flooding.
• This approach induce a description of the basin as a set of areas
  • the one are in a lack of protection (red)
  • the other one are "underflooded" (green).
• Relevant decision board can decide
  • to freeze some areas for them not to turn red soon, to modify land use, or to spatially modify the alea pattern with minimal river works, turning areas red to green at the "hydrological expenses" of green
• The alea / vulnerability approach can be done
  • via administrative measures, or
  • via local negociations o including payment to insurance companies
• according to the cultural habits of each community.

definitively lacking data

• rain known via
  • rain gauges select 400 cm2 in 100 km2
  • weather radar
• spatial pattern, but little quantitative consistency o potential evapotranspiration known via
  • observed meteorological estimation of control factors (temperature, wind), at 100 km grid size

definitively lacking data of real evapotranspiration known

• only via water balance estimation at the field or basin scale o discharge
• known at 15 % in some gaging stations (500 working stations in France).
• include non registered man-made perturbations that make the assessment of the intrinsic behaviour of the catchment very difficult

the hierarchy of processes is unstable

• a process can easily take precedence on an other because of
  • the quasi-systematic non linearity of processes
  • their sensitivity to the initial conditions
• effect of water contents
• effect of soil structure

a catchment......

• can have a behaviour that is completely dominated by some usually neglected process
• as a behaviour that is not uniquely determine by the contents, but also by their spatial organisation
• comparison with a recipe
  • we know the taste of each ingredient.
  • we can NOT predict the taste of the meal

examples of atypical conditions :

• Zebra bush in sahelian regions
• Mulch
• Snow redistribution by the wind
• Groundwater sustained rivers
• Man-made linear patterns in landscape

a built-in link with other specialities

• Soil physics and plant physiology
• Water quality, hydrobiology
• River geomorphology
• Human and social sciences
• Management and economy, law, politics

scientific reasons to build models

• Blackboard tool
  • formalisation of concepts
  • possible formal checking
  • knowledge and concepts
• Data interpretation
• Behavioural simulation
  • explicitation of non-obvious structure effects

operational reasons to build models

• answering specialized questions
  • assessing impacts of land-use change
• testing general management strategies

scope of the model ?

• which area ?
• which level of detail ?
  • are the details useful ?
  • will we be able to gather the details ?
• which time scope - season ?
  • duration ?
  • climate and social scenarios ?
• which hydrologically related features do we need ?
  • floods ; water quantity ; water quality ; hydrobiology ; river geomorphology ; water uses ; land use
• choice of independant and dependant features ?

some critical points in hydrological modelisation

• Assessing the dominant processes
  • Is there a link to what I am interested to ?
• Choosing time and space scales
• Choosing a topology
• Is an object oriented approach usefull ?
  • How to separate objects ?
  • How to specify the relation between objects ?

models relationship to causality

• Deterministic models : deductive models
• Statistical models : inductive models
  • probabilistic models o directly on distributions
  • stochastic models
• yielding time-series as output

lumped models

• boxes flowing the ones into the others through pipes...
• need for calibration
• useful as reference catchments in applications involving reference catchments
  • detection of changes

steps in elaborating a lumped, conceptual hydrologic model

• identification (which structure ?)
• calibration (value of parameters)
• validation (check)
• documentation of limits
  • physical limits
  • numerical limits

distributed models

• according to a regular grid
  • an old-fashioned, quite efficient way
• according to a dominant process-based grid
  • slopes and contours
• according to an homogeneous area concept
  • valid only in man-made landscape
  • terrific topology
• a general tool would need the tree forms to be easily mixed!

• mostly for regular grid distributed models
• physical parameters unknown and spatially variable at the sub-grid resolution
  • effective parameters approach :
• equations are kept same as in the detailed scale, but with (possibly other) numerical values that account for macroscopic scale behaviour
• parametrization approach
  • given a scheme of what the subgrid variability is, a stochastic approach derives a set of macroscopically suitable equations that have a form that is not the same as the one of the small scale o inverse approach - parameters are estimated backward from overall behaviour of the catchment
• integrated measurement
  • remote sensors are supposedly able to evaluate some characteristic parameters of the surface (moisture, rugosity, slopeÖ) directly at a scale that is suitable for distributed modeling

explicit physics and parametrisation

• part of explicit physics quite modest.
• unresolved part
  • accounted for via behavioural routines
  • tend to be the core of models (not just in well localized "parametrisation boxes").
• models who clame to be deterministic (for they are distributed) may be completely behavioural when one consider the scheme implemented at the cell size.

basic aspects of hydrology

the aim of hydrology:

To determine how much water will be in a given location and condition. This is a more complex issue than it at first might seem; for example, as local mayor may want to address local flooding, but this is not a purely local issue, and the areas up and downstream need to be considered: there is the need then to 'zoom out'.

The hydrological cycle: Is a continuum, broken by the observer into

• storages
  • water bodies
  • with possible internal evolutionary laws
• water fluxes
  • inside or between water bodies
  • associated to hydrological processes

The main freshwater storages are: Ranked here by increasing time constant - atmosphere

• soil moisture (non saturated area)
• rivers
• snowpack
• lakes ; reservoirs
• groundwater (saturated area)
• icepack

The main freshwater fluxes include:

• Precipitation
• (actual) Evapotranspiration
• Infiltration and seepage (= ex filtration)
• Runoff (on slopes)
• Discharge (in rivers)

hydrological processes:

• Water fluxes are linked to hydrological processes:
  • not only fluxes between water bodies
  • also internal evolution of water bodies
• A process is an elementary behaviour that can described as a whole, and whose level of formalisation may vary. It is under control of various factors

sample processes

Sample process: runoff formation

• according to Horton, runoff occurs where and when rain rate exceeds infiltration capacity
• according to Capus, Hewlett, Beven, runoff occurs where and when rain falls on saturated areas
• importance of the soil structure

Sample process: runoff collection to discharge

• Overland flow (on slopes)
  • Gullies, connectivity topics
  • Importance of relative location of land use
  • Importance of subrogate features of land use (direction of ploughing)

Sample process: underground flow

• The continuous model
  • unsatured zone : the Richards equation
  • satured zone : the Darcy equation
• local formula
• integrated form for alluvial aquifers
• integrated form for constrained aquifer
  • The problem of parameters estimation
• importance of K(q, x, y, z) (a tensor)
• Preferential pathes
  • biological macropores
  • pipes
  • roots
• Impervious layers
  • bottom of ploughing area

catchment

An key hydrological object is the catchment (= the basin)

• An outlet
• The river network upstream
• Slopes - both side of the rivers
  • up to the water divides
• Includes
  • surface and subsurface storages in relation to the river

• the best possible system to study as far as geophysical fluxes as considered
  • one input (rain, other atmospheric conditions)
  • one output (discharge at the outlet)
• the best possible unit for effective management
  • what I do here is my problem

• A fully explicit, exhaustive description is impossible because of
  • the fractal nature of the river network
  • the fractal nature of the topography
  • the partially unreachable description of the under ground
  • the unsteady character of the topography and soil properties at detailed scale
• The catchment is
  • a point (the outlet)
  • a set of lines (the river network)
  • an area (interacting with the atmosphere),
  • a volume (including the underground). o Implementation of such an object in a GIS is not straightforward.
• The definition of a catchment is outlet dependent.
• Two gauging stations define either nested or non-nested catchments
• Data out of many catchments are part of a data hierarchy that must sometimes be considered explicitly (discharge mapping).
• Some problems seem point oriented...
  • ´ how can I reduced floods here -a Ö but must be handled considering all the processes - upstream (causes) and
  • downstream (consequencies of options to take). Often we have to ´ zoom out -a to have a grasp at the problem as a whole.
• It is usually not an administrative division
• The concept may break down - karstic areas - flat, human dominated areas

hydrological monographies

• A balanced description of a catchment (hydrological monography) can be very interesting.
• It will not solve all possible and unexpected questions. A problem oriented description of a catchment
• needs a variety of choices to be done
  • selecting the processes relevant to the problem
  • the scale of the features to explicitely take into account.
  • the time to be considered
• the abduction of non-relevant details has to remain in mind.

Production and transfer functions

• ´ Production -a
  • relates the gross precipitation over the catchment to the net precipitation that is to flow through the outlet.
  • non-flowing water is only considered as a soil moisture controlling factor, influencing the soil behaviour under further rains.
• ´ Transfer -a
  • relates the produced ´ net precipitation -a to the discharge.

About this scheme

• It is common choice to
  • - upload the production function with all the non-linearity of the rain-discharge transformation.
  • -consider the transfert function as linear.
• This approximation may be valid for heavy rains

conceptual approaches to the transfer aspects

• Unit Hydrograph (Sherman, 1932)
  • the transfer function is assumed linear.
  • the structure of the non-linear production function remains author-dependant.
  • parameters for both parts are identified from a joint pair of long rainfall/discharge time series.
• Geomorphologic Unit Hydrograph :
  • an improvement from the previous approach
  • the shape of the unit hydrograph is related to distances and slopes along the runoff pathways from the catchment to the outlet
  • this gives clearer constraints to what the production function can be

limits to these approaches

• Isotopes evaluations show that most of the water of the flood has been in the soil long before the begin of the rain.

• Continuity equation
• Dynamic equation
  • Head
• potential energy + kinetic energy
  • Head losses o along the stream (energy loss in turbulence, interactions between the water and the reach)
• localised (in hydraulic jumps from torrential to fluvial conditions

various levels of description for hydraulic transfer

• in general, PDE equations
• 3D equations (Navier Stokes)
  • small scale studies like geomorphology, flow around a bridge
• 2D equations (BarrÈ Saint-Venant)
  • where overland flow is most relevant : dam breakes, flooding of broader areas with non negligible speed in the flooded part
• 1D+storages (Barre Saint-Venant) :
  • where the flooded area is broken in independent storages, where speed is negligible
• 1D (Barre Saint-Venant) :
  • where streamflow is concentrated in the minor riverbed (no flooding).
  • including dam breaks, working spillways, moving hydraulic jumps, ...
• Simplified 1D equations :
  • Diffusive wave approximation :
• flood diffusion in gentle, sub-horizontal rivers
  • Cinematic wave approximation :
• flood propagation in steep rivers or lateral slopes
• 1D, steady-state approximation :
  • if time variations are negligible. Mostly broad, gentle rivers,
• a important step for text-books in hydraulics (clear, intuitive relation of results to energetic consideration and limits)
• 1D, uniform approximation :
  • to be considered only in regular, chenalized reaches

• To predict floods, or to assess flood hazard?
  • To predict o Given a current stage of water and observed or predicted rain, guess the shape, time of arrival and water stage to occur in the next future at the interest point.
  • To assess o Given a observed discharge time-serie, give probability of a given flood characteristic (peak flow, duration, volume) to be over-seeded

• who
  • civil servants; river authorities; mayors; meteorologists; hydrologists
• how
  • real time data collection
  • quick data processing, mostly empirical models or analogues
  • 365 days, 24 hour communication system to people
• what
  • technical choice of a flood index to predict, level of confidence
• to who
  • police, municipality representatives, everybody ?
• what to say
  • how clear the warning messages ?
  • readiness to cooperate ?

a personal interpretation

• some rivers have long time constants
  • gentle rain, so progressive saturation ; broad basins, so long hydraulic transfers
• some rivers have short time constants
  • steep, small catchments ; convective storms.
• this
  • enable different kind of human measures
  • induced an "hydrology of flash floods" to exist
• but hydrology is one!

flood management approaches

• flood
  • is a natural event
  • can be characterised as an random event => alea
• flooding
  • can yield damages
  • this depends on the sensitivity of land use vulnerability

the damage approach : principle

• considers vulnerability as the cost of damages
• to minimise by - protective measures (levees),
  • storing or evacuating waters via various works,
• as far as monetary evaluation proves efficient.

the damage approach : drawback

• Due to ...
  • the probabilistic nature of events,
  • the short memory of human beings,
  • teleconnections of local actions and basin-wide effects,
• spontaneous local management exhibits a drift towards heavy works that appears to be unsustainable at the basin scale (spiral of corrective measures).

the alea / vulnerability approach

• vulnerability of each type of landuse is a socially determined, possibly negociated acceptance for flooding
• some areas, like marshes, may have a positive demand for flooding.
• This approach induce a description of the basin as a set of areas
  • the one are in a lack of protection (red)
  • the other one are "underflooded" (green).
• Relevant decision board can decide
  • to freeze some areas for them not to turn red soon, to modify land use, or to spatially modify the alea pattern with minimal river works, turning areas red to green at the "hydrological expenses" of green
• The alea / vulnerability approach can be done
  • via administrative measures, or
  • via local negociations o including payment to insurance companies
• according to the cultural habits of each community.

definitively lacking data

• rain known via
  • rain gauges select 400 cm2 in 100 km2
  • weather radar
• spatial pattern, but little quantitative consistency o potential evapotranspiration known via
  • observed meteorological estimation of control factors (temperature, wind), at 100 km grid size

definitively lacking data of real evapotranspiration known

• only via water balance estimation at the field or basin scale o discharge
• known at 15 % in some gaging stations (500 working stations in France).
• include non registered man-made perturbations that make the assessment of the intrinsic behaviour of the catchment very difficult

the hierarchy of processes is unstable

• a process can easily take precedence on an other because of
  • the quasi-systematic non linearity of processes
  • their sensitivity to the initial conditions
• effect of water contents
• effect of soil structure

a catchment......

• can have a behaviour that is completely dominated by some usually neglected process
• as a behaviour that is not uniquely determine by the contents, but also by their spatial organisation
• comparison with a recipe
  • we know the taste of each ingredient.
  • we can NOT predict the taste of the meal

examples of atypical conditions :

• Zebra bush in sahelian regions
• Mulch
• Snow redistribution by the wind
• Groundwater sustained rivers
• Man-made linear patterns in landscape

a built-in link with other specialities

• Soil physics and plant physiology
• Water quality, hydrobiology
• River geomorphology
• Human and social sciences
• Management and economy, law, politics

scientific reasons to build models

• Blackboard tool
  • formalisation of concepts
  • possible formal checking
  • knowledge and concepts
• Data interpretation
• Behavioural simulation
  • explicitation of non-obvious structure effects

operational reasons to build models

• answering specialized questions
  • assessing impacts of land-use change
• testing general management strategies

scope of the model ?

• which area ?
• which level of detail ?
  • are the details useful ?
  • will we be able to gather the details ?
• which time scope - season ?
  • duration ?
  • climate and social scenarios ?
• which hydrologically related features do we need ?
  • floods ; water quantity ; water quality ; hydrobiology ; river geomorphology ; water uses ; land use
• choice of independant and dependant features ?

some critical points in hydrological modelisation

• Assessing the dominant processes
  • Is there a link to what I am interested to ?
• Choosing time and space scales
• Choosing a topology
• Is an object oriented approach usefull ?
  • How to separate objects ?
  • How to specify the relation between objects ?

models relationship to causality

• Deterministic models : deductive models
• Statistical models : inductive models
  • probabilistic models o directly on distributions
  • stochastic models
• yielding time-series as output

lumped models

• boxes flowing the ones into the others through pipes...
• need for calibration
• useful as reference catchments in applications involving reference catchments
  • detection of changes

steps in elaborating a lumped, conceptual hydrologic model

• identification (which structure ?)
• calibration (value of parameters)
• validation (check)
• documentation of limits
  • physical limits
  • numerical limits

distributed models

• according to a regular grid
  • an old-fashioned, quite efficient way
• according to a dominant process-based grid
  • slopes and contours
• according to an homogeneous area concept
  • valid only in man-made landscape
  • terrific topology
• a general tool would need the tree forms to be easily mixed!

• mostly for regular grid distributed models
• physical parameters unknown and spatially variable at the sub-grid resolution
  • effective parameters approach :
• equations are kept same as in the detailed scale, but with (possibly other) numerical values that account for macroscopic scale behaviour
• parametrization approach
  • given a scheme of what the subgrid variability is, a stochastic approach derives a set of macroscopically suitable equations that have a form that is not the same as the one of the small scale o inverse approach - parameters are estimated backward from overall behaviour of the catchment
• integrated measurement
  • remote sensors are supposedly able to evaluate some characteristic parameters of the surface (moisture, rugosity, slopeÖ) directly at a scale that is suitable for distributed modeling

explicit physics and parametrisation

• part of explicit physics quite modest.
• unresolved part
  • accounted for via behavioural routines
  • tend to be the core of models (not just in well localized "parametrisation boxes").
• models who clame to be deterministic (for they are distributed) may be completely behavioural when one consider the scheme implemented at the cell size.

examples of hydrological models

• square grid, physically based
  • SHE
• contour and slope grid, physically based
  • TOPOG
• square grid, conceptual
  • Stanford IV, Cequeau, ModCou o lumped, conceptual
  • CREC, GR4J, Gardenia
• semi-lumped, specialised to saturation runoff : Topmodel

the Chestnut Valley: an example

Background

• Privas, ArdËche dept, France
• Key industry : chesnut processing (Christmas, etc.)
• On the OuvËze river, a tributary to the Rhne
• Some agricultural opportunities in the valley, downstream from Privas

The state today of the chestnut valley

• two tributaries of the Ouveze are used for providing water to the chestnut industry.
• Water shortages in Privas
• Ouveze dry off in summer in Privas
• Ouveze is merely chesnut waste donwstream from Privas ; biology near to down to the Rhne.
• Agriculture does not really start, because lacking water

basic aspects of hydrology

the aim of hydrology:

To determine how much water will be in a given location and condition. This is a more complex issue than it at first might seem; for example, as local mayor may want to address local flooding, but this is not a purely local issue, and the areas up and downstream need to be considered: there is the need then to 'zoom out'.

The hydrological cycle: Is a continuum, broken by the observer into

• storages
  • water bodies
  • with possible internal evolutionary laws
• water fluxes
  • inside or between water bodies
  • associated to hydrological processes

The main freshwater storages are: Ranked here by increasing time constant - atmosphere

• soil moisture (non saturated area)
• rivers
• snowpack
• lakes ; reservoirs
• groundwater (saturated area)
• icepack

The main freshwater fluxes include:

• Precipitation
• (actual) Evapotranspiration
• Infiltration and seepage (= ex filtration)
• Runoff (on slopes)
• Discharge (in rivers)

hydrological processes:

• Water fluxes are linked to hydrological processes:
  • not only fluxes between water bodies
  • also internal evolution of water bodies
• A process is an elementary behaviour that can described as a whole, and whose level of formalisation may vary. It is under control of various factors

sample processes

Sample process: runoff formation

• according to Horton, runoff occurs where and when rain rate exceeds infiltration capacity
• according to Capus, Hewlett, Beven, runoff occurs where and when rain falls on saturated areas
• importance of the soil structure

Sample process: runoff collection to discharge

• Overland flow (on slopes)
  • Gullies, connectivity topics
  • Importance of relative location of land use
  • Importance of subrogate features of land use (direction of ploughing)

Sample process: underground flow

• The continuous model
  • unsatured zone : the Richards equation
  • satured zone : the Darcy equation
• local formula
• integrated form for alluvial aquifers
• integrated form for constrained aquifer
  • The problem of parameters estimation
• importance of K(q, x, y, z) (a tensor)
• Preferential pathes
  • biological macropores
  • pipes
  • roots
• Impervious layers
  • bottom of ploughing area

catchment

An key hydrological object is the catchment (= the basin)

• An outlet
• The river network upstream
• Slopes - both side of the rivers
  • up to the water divides
• Includes
  • surface and subsurface storages in relation to the river

• the best possible system to study as far as geophysical fluxes as considered
  • one input (rain, other atmospheric conditions)
  • one output (discharge at the outlet)
• the best possible unit for effective management
  • what I do here is my problem

• A fully explicit, exhaustive description is impossible because of
  • the fractal nature of the river network
  • the fractal nature of the topography
  • the partially unreachable description of the under ground
  • the unsteady character of the topography and soil properties at detailed scale
• The catchment is
  • a point (the outlet)
  • a set of lines (the river network)
  • an area (interacting with the atmosphere),
  • a volume (including the underground). o Implementation of such an object in a GIS is not straightforward.
• The definition of a catchment is outlet dependent.
• Two gauging stations define either nested or non-nested catchments
• Data out of many catchments are part of a data hierarchy that must sometimes be considered explicitly (discharge mapping).
• Some problems seem point oriented...
  • ´ how can I reduced floods here -a Ö but must be handled considering all the processes - upstream (causes) and
  • downstream (consequencies of options to take). Often we have to ´ zoom out -a to have a grasp at the problem as a whole.
• It is usually not an administrative division
• The concept may break down - karstic areas - flat, human dominated areas

hydrological monographies

• A balanced description of a catchment (hydrological monography) can be very interesting.
• It will not solve all possible and unexpected questions. A problem oriented description of a catchment
• needs a variety of choices to be done
  • selecting the processes relevant to the problem
  • the scale of the features to explicitely take into account.
  • the time to be considered
• the abduction of non-relevant details has to remain in mind.

Production and transfer functions

• ´ Production -a
  • relates the gross precipitation over the catchment to the net precipitation that is to flow through the outlet.
  • non-flowing water is only considered as a soil moisture controlling factor, influencing the soil behaviour under further rains.
• ´ Transfer -a
  • relates the produced ´ net precipitation -a to the discharge.

About this scheme

• It is common choice to
  • - upload the production function with all the non-linearity of the rain-discharge transformation.
  • -consider the transfert function as linear.
• This approximation may be valid for heavy rains

conceptual approaches to the transfer aspects

• Unit Hydrograph (Sherman, 1932)
  • the transfer function is assumed linear.
  • the structure of the non-linear production function remains author-dependant.
  • parameters for both parts are identified from a joint pair of long rainfall/discharge time series.
• Geomorphologic Unit Hydrograph :
  • an improvement from the previous approach
  • the shape of the unit hydrograph is related to distances and slopes along the runoff pathways from the catchment to the outlet
  • this gives clearer constraints to what the production function can be

limits to these approaches

• Isotopes evaluations show that most of the water of the flood has been in the soil long before the begin of the rain.

• Continuity equation
• Dynamic equation
  • Head
• potential energy + kinetic energy
  • Head losses o along the stream (energy loss in turbulence, interactions between the water and the reach)
• localised (in hydraulic jumps from torrential to fluvial conditions

various levels of description for hydraulic transfer

• in general, PDE equations
• 3D equations (Navier Stokes)
  • small scale studies like geomorphology, flow around a bridge
• 2D equations (BarrÈ Saint-Venant)
  • where overland flow is most relevant : dam breakes, flooding of broader areas with non negligible speed in the flooded part
• 1D+storages (Barre Saint-Venant) :
  • where the flooded area is broken in independent storages, where speed is negligible
• 1D (Barre Saint-Venant) :
  • where streamflow is concentrated in the minor riverbed (no flooding).
  • including dam breaks, working spillways, moving hydraulic jumps, ...
• Simplified 1D equations :
  • Diffusive wave approximation :
• flood diffusion in gentle, sub-horizontal rivers
  • Cinematic wave approximation :
• flood propagation in steep rivers or lateral slopes
• 1D, steady-state approximation :
  • if time variations are negligible. Mostly broad, gentle rivers,
• a important step for text-books in hydraulics (clear, intuitive relation of results to energetic consideration and limits)
• 1D, uniform approximation :
  • to be considered only in regular, chenalized reaches

• To predict floods, or to assess flood hazard?
  • To predict o Given a current stage of water and observed or predicted rain, guess the shape, time of arrival and water stage to occur in the next future at the interest point.
  • To assess o Given a observed discharge time-serie, give probability of a given flood characteristic (peak flow, duration, volume) to be over-seeded

• who
  • civil servants; river authorities; mayors; meteorologists; hydrologists
• how
  • real time data collection
  • quick data processing, mostly empirical models or analogues
  • 365 days, 24 hour communication system to people
• what
  • technical choice of a flood index to predict, level of confidence
• to who
  • police, municipality representatives, everybody ?
• what to say
  • how clear the warning messages ?
  • readiness to cooperate ?

a personal interpretation

• some rivers have long time constants
  • gentle rain, so progressive saturation ; broad basins, so long hydraulic transfers
• some rivers have short time constants
  • steep, small catchments ; convective storms.
• this
  • enable different kind of human measures
  • induced an "hydrology of flash floods" to exist
• but hydrology is one!

flood management approaches

• flood
  • is a natural event
  • can be characterised as an random event => alea
• flooding
  • can yield damages
  • this depends on the sensitivity of land use vulnerability

the damage approach : principle

• considers vulnerability as the cost of damages
• to minimise by - protective measures (levees),
  • storing or evacuating waters via various works,
• as far as monetary evaluation proves efficient.

the damage approach : drawback

• Due to ...
  • the probabilistic nature of events,
  • the short memory of human beings,
  • teleconnections of local actions and basin-wide effects,
• spontaneous local management exhibits a drift towards heavy works that appears to be unsustainable at the basin scale (spiral of corrective measures).

the alea / vulnerability approach

• vulnerability of each type of landuse is a socially determined, possibly negociated acceptance for flooding
• some areas, like marshes, may have a positive demand for flooding.
• This approach induce a description of the basin as a set of areas
  • the one are in a lack of protection (red)
  • the other one are "underflooded" (green).
• Relevant decision board can decide
  • to freeze some areas for them not to turn red soon, to modify land use, or to spatially modify the alea pattern with minimal river works, turning areas red to green at the "hydrological expenses" of green
• The alea / vulnerability approach can be done
  • via administrative measures, or
  • via local negociations o including payment to insurance companies
• according to the cultural habits of each community.

definitively lacking data

• rain known via
  • rain gauges select 400 cm2 in 100 km2
  • weather radar
• spatial pattern, but little quantitative consistency o potential evapotranspiration known via
  • observed meteorological estimation of control factors (temperature, wind), at 100 km grid size

definitively lacking data of real evapotranspiration known

• only via water balance estimation at the field or basin scale o discharge
• known at 15 % in some gaging stations (500 working stations in France).
• include non registered man-made perturbations that make the assessment of the intrinsic behaviour of the catchment very difficult

the hierarchy of processes is unstable

• a process can easily take precedence on an other because of
  • the quasi-systematic non linearity of processes
  • their sensitivity to the initial conditions
• effect of water contents
• effect of soil structure

a catchment......

• can have a behaviour that is completely dominated by some usually neglected process
• as a behaviour that is not uniquely determine by the contents, but also by their spatial organisation
• comparison with a recipe
  • we know the taste of each ingredient.
  • we can NOT predict the taste of the meal

examples of atypical conditions :

• Zebra bush in sahelian regions
• Mulch
• Snow redistribution by the wind
• Groundwater sustained rivers
• Man-made linear patterns in landscape

a built-in link with other specialities

• Soil physics and plant physiology
• Water quality, hydrobiology
• River geomorphology
• Human and social sciences
• Management and economy, law, politics

scientific reasons to build models

• Blackboard tool
  • formalisation of concepts
  • possible formal checking
  • knowledge and concepts
• Data interpretation
• Behavioural simulation
  • explicitation of non-obvious structure effects

operational reasons to build models

• answering specialized questions
  • assessing impacts of land-use change
• testing general management strategies

scope of the model ?

• which area ?
• which level of detail ?
  • are the details useful ?
  • will we be able to gather the details ?
• which time scope - season ?
  • duration ?
  • climate and social scenarios ?
• which hydrologically related features do we need ?
  • floods ; water quantity ; water quality ; hydrobiology ; river geomorphology ; water uses ; land use
• choice of independant and dependant features ?

some critical points in hydrological modelisation

• Assessing the dominant processes
  • Is there a link to what I am interested to ?
• Choosing time and space scales
• Choosing a topology
• Is an object oriented approach usefull ?
  • How to separate objects ?
  • How to specify the relation between objects ?

models relationship to causality

• Deterministic models : deductive models
• Statistical models : inductive models
  • probabilistic models o directly on distributions
  • stochastic models
• yielding time-series as output

lumped models

• boxes flowing the ones into the others through pipes...
• need for calibration
• useful as reference catchments in applications involving reference catchments
  • detection of changes

steps in elaborating a lumped, conceptual hydrologic model

• identification (which structure ?)
• calibration (value of parameters)
• validation (check)
• documentation of limits
  • physical limits
  • numerical limits

distributed models

• according to a regular grid
  • an old-fashioned, quite efficient way
• according to a dominant process-based grid
  • slopes and contours
• according to an homogeneous area concept
  • valid only in man-made landscape
  • terrific topology
• a general tool would need the tree forms to be easily mixed!

• mostly for regular grid distributed models
• physical parameters unknown and spatially variable at the sub-grid resolution
  • effective parameters approach :
• equations are kept same as in the detailed scale, but with (possibly other) numerical values that account for macroscopic scale behaviour
• parametrization approach
  • given a scheme of what the subgrid variability is, a stochastic approach derives a set of macroscopically suitable equations that have a form that is not the same as the one of the small scale o inverse approach - parameters are estimated backward from overall behaviour of the catchment
• integrated measurement
  • remote sensors are supposedly able to evaluate some characteristic parameters of the surface (moisture, rugosity, slopeÖ) directly at a scale that is suitable for distributed modeling

explicit physics and parametrisation

• part of explicit physics quite modest.
• unresolved part
  • accounted for via behavioural routines
  • tend to be the core of models (not just in well localized "parametrisation boxes").
• models who clame to be deterministic (for they are distributed) may be completely behavioural when one consider the scheme implemented at the cell size.

examples of hydrological models

• square grid, physically based
  • SHE
• contour and slope grid, physically based
  • TOPOG
• square grid, conceptual
  • Stanford IV, Cequeau, ModCou o lumped, conceptual
  • CREC, GR4J, Gardenia
• semi-lumped, specialised to saturation runoff : Topmodel

the Chestnut Valley: an example

Background

• Privas, ArdËche dept, France
• Key industry : chesnut processing (Christmas, etc.)
• On the OuvËze river, a tributary to the Rhne
• Some agricultural opportunities in the valley, downstream from Privas

The state today of the chestnut valley

• two tributaries of the Ouveze are used for providing water to the chestnut industry.
• Water shortages in Privas
• Ouveze dry off in summer in Privas
• Ouveze is merely chesnut waste donwstream from Privas ; biology near to down to the Rhne.
• Agriculture does not really start, because lacking water

Spontaneous sectorial remediation projects

• for problems in Privas
  • building dams on the tributaries for an enhancement of water availability in Privas ; maybe, to sustain summer discharge of the Ouveze
• for agriculture
  • building a irrigation pipe from the Rhne
• An idea for integrated management
  • irrigation pipe to go up to Privas
  • chestnut waste to be diverted to the agricultural areas
• Expected benefits
  • abundant water to the industry and inhabitants o dam project can be forgotten
  • river will biologically recover
• A need o evaluate this and others scheme quickly

A point raised it that small events may be forgotten by policy makers. If you are not experiencing floods, it is difficult to believe that you are at risk of a big flood. In recently developed areas, there is no memory of what risk can be, and in the chain of decision making, people are not aware of what the risks are. The absence of risk through engineering control means that people do not think what will happen. But by thinking that small scale risk is decreasing, this means that there will be less anxiety about big risks, which may still present a danger.

Using physical catchment as the scale of measurement has been brought into question Problems of demand for water from different sources are relevant here. For example, if tourism or agriculture take water not from their own catchment. By not having an integrated framework, it may be that stakeholders may achieve their objectives more easily. The Chestnut Valley is an example of wherelocal action is not the answer. The local suggestion would be to build a damm, but analysis suggests other options, such as a pipe to bring water in from other areas.

Spontaneous sectorial remediation projects

• for problems in Privas
  • building dams on the tributaries for an enhancement of water availability in Privas ; maybe, to sustain summer discharge of the Ouveze
• for agriculture
  • building a irrigation pipe from the Rhne
• An idea for integrated management
  • irrigation pipe to go up to Privas
  • chestnut waste to be diverted to the agricultural areas
• Expected benefits
  • abundant water to the industry and inhabitants o dam project can be forgotten
  • river will biologically recover
• A need o evaluate this and others scheme quickly

A point raised it that small events may be forgotten by policy makers. If you are not experiencing floods, it is difficult to believe that you are at risk of a big flood. In recently developed areas, there is no memory of what risk can be, and in the chain of decision making, people are not aware of what the risks are. The absence of risk through engineering control means that people do not think what will happen. But by thinking that small scale risk is decreasing, this means that there will be less anxiety about big risks, which may still present a danger.

Using physical catchment as the scale of measurement has been brought into question Problems of demand for water from different sources are relevant here. For example, if tourism or agriculture take water not from their own catchment. By not having an integrated framework, it may be that stakeholders may achieve their objectives more easily. The Chestnut Valley is an example of wherelocal action is not the answer. The local suggestion would be to build a damm, but analysis suggests other options, such as a pipe to bring water in from other areas.

examples of hydrological models

• square grid, physically based
  • SHE
• contour and slope grid, physically based
  • TOPOG
• square grid, conceptual
  • Stanford IV, Cequeau, ModCou o lumped, conceptual
  • CREC, GR4J, Gardenia
• semi-lumped, specialised to saturation runoff : Topmodel

the Chestnut Valley: an example

Background

• Privas, ArdËche dept, France
• Key industry : chesnut processing (Christmas, etc.)
• On the OuvËze river, a tributary to the Rhne
• Some agricultural opportunities in the valley, downstream from Privas

The state today of the chestnut valley

• two tributaries of the Ouveze are used for providing water to the chestnut industry.
• Water shortages in Privas
• Ouveze dry off in summer in Privas
• Ouveze is merely chesnut waste donwstream from Privas ; biology near to down to the Rhne.
• Agriculture does not really start, because lacking water

Spontaneous sectorial remediation projects

• for problems in Privas
  • building dams on the tributaries for an enhancement of water availability in Privas ; maybe, to sustain summer discharge of the Ouveze
• for agriculture
  • building a irrigation pipe from the Rhne
• An idea for integrated management
  • irrigation pipe to go up to Privas
  • chestnut waste to be diverted to the agricultural areas
• Expected benefits
  • abundant water to the industry and inhabitants o dam project can be forgotten
  • river will biologically recover
• A need o evaluate this and others scheme quickly

A point raised it that small events may be forgotten by policy makers. If you are not experiencing floods, it is difficult to believe that you are at risk of a big flood. In recently developed areas, there is no memory of what risk can be, and in the chain of decision making, people are not aware of what the risks are. The absence of risk through engineering control means that people do not think what will happen. But by thinking that small scale risk is decreasing, this means that there will be less anxiety about big risks, which may still present a danger.

Using physical catchment as the scale of measurement has been brought into question Problems of demand for water from different sources are relevant here. For example, if tourism or agriculture take water not from their own catchment. By not having an integrated framework, it may be that stakeholders may achieve their objectives more easily. The Chestnut Valley is an example of wherelocal action is not the answer. The local suggestion would be to build a damm, but analysis suggests other options, such as a pipe to bring water in from other areas.