User Tools

Site Tools


start:hype_model_description:hype_human_water

Water management

Constructed wetlands

For an overview of basic assumptions and explanation of variables see the Basic assumptions section in the Rivers and lakes chapter.

The wetlands that are simulated are small artificial ponds. They have an area and depth, but their area is not taken into account in terms of precipitation and evaporation. The water flow passes through the wetlands without being affected, so it's just as nutrient traps that the wetland model is significant. There are two types of wetlands, just as for the rivers and lakes. They are situated before the river in the calculation scheme. The local wetland (lrwet) receives a share of the local runoff the rest passes by unaffected. Wetlands in main rivers (mrwet) receive a portion of the flow in the main river and the rest passes unaffected.

Wetland nutrient processes

In wetlands retention of inorganic nitrogen is modelled. For total phosphorus retention and production of TP are modelled. The rates of these processes are constant (teta=1.2, tkoeff=20, inpar=2.3, sedpar= 0.09, and uptpar=0.1). The retention is limited to 99.9% of the substance in the wetland. The retention (retIN, retTP, g/d) depends on the rate parameter, the concentration in the wetland, wetland area, and for inorganic nitrogen also on 5-day-mean temperature. The production (prodTP, g/d) depends on a rate parameter, the concentration of the inflow to the wetland, wetland area, and a temperature function.

retIN = inpar * INconc * temp5 * area / 1000.
retTP = sedpar * TPconc * area
prodTP = uptpar * TPin * area * teta**(temp30-tcoeff)

Modules (file) Procedures
npc_surfacewater_processes (npc_sw_proc.f90) calculate_river_wetland
calculate_wetland_np

Irrigation

General principles

Irrigation constitutes a key water management activity in many parts of the world. Therefore, the HYPE model has a routine to simulate irrigation. The representation of irrigation in the model is based on a set of principles. Firstly, the irrigation water demand is assessed. Subsequently, the demanded water is withdrawn from the defined irrigation water sources. HYPE can either withdraw water from defined sub-basins in the model domain (subject to availability), or from unlimited sources outside the domain. Finally, the withdrawn water is applied onto the classes from which the demand originated. In addition, water losses between demand, withdrawal, and application are taken into account (for withdrawals within the model domain).

A class is irrigated if the crop type associated with it is irrigated (defined in GeoClass.txt). A crop is irrigated if the irrigation input variables in the CropData.txt file are defined and non-zero (plantday, lengthini, kcbini, lengthdev, lengthmid, kcbmid, lengthlate, kcbend, dlref). Irrigation also requires appropriate values in the MgmtData.txt file (gw_part, regsrcid, irrdam, region_eff, local_eff, demandtype) and the par.txt file (pirrs, pirrg, sswcorr etc.). See the FileDescription document for more details on each file and each parameter.

Irrigation water demand

The irrigation water demand (W_{I,D}) is calculated each day for each irrigated class (j) at the end of the soil water balance calculations. Two approaches to calculate W_{I,D,j} are implemented in HYPE, one for submerged crops (e.g. paddy rice) and one for non-submerged crops. The input variables imm_start and imm_end in CropData.txt define (1) the beginning and end of the submerged season, and (2) if crops are submerged or not (zero is interpreted as a non-submerged crop).

Non-submerged crops

For non-submerged crops, the calculations are based on the FAO-56 crop coefficient methods (Allen et al., 1998). The dual crop coefficient method is used because it is more specific than the single crop coefficient method, and more suitable for daily water balance models. Since transpiration is of primary interest in estimating crop water demand, the irrigation routine focuses on estimating potential transpiration (T_P) with the basal crop coefficient (K_{CB}) and the reference potential crop evapotranspiration (ET_0):

T_P=K_{CB}×ET_0

ET0 follows the dynamics described above (ET_0 = epot here following Wisser et al. (2008)). KCB depends on crop type and phenological stage, which is defined in CropData.txt. KCB is constant during the initial development stage, then increases linearly during the development stage until it reaches the mid-season stage during which it is again constant. Finally, K_{CB} decreases linearly from the end of the mid-season stage until the end of the season. The dynamics of ET_0 and K_{CB} produces a dynamic T_P profile (Figure 1). Allen et al. (1998) provide indicative values for K_{CB} (cf. their Table 17).

Figure 1 Illustration of ET_0, T_P, K_CB, and SSW for a typical maize crop on a medium coarse soil in southern Europe. Key input variables to define the K_CB profile are shown in blue.

On any given day, the model first calculates whether irrigation is needed, and then the amount required. The irrigation need is assessed by comparing the current soil water content (H) with a dynamic irrigation threshold (S_{SW}, the soil water stress threshold):

if H <(S_{SW}×P_{SSWCORR}×AWC) right irrigate

H is the plant-available soil moisture (i.e. soil water above wcwp1 and wcwp2 in soil layers 1 and 2 respectively). AWC is the maximum plant-available water content in soil layers 1 and 2 (i.e. the sum of fc1 and fc2). S_{SW} is a fraction of AWC (defined upwards from wcwp). Below S_{SW} the crop experiences water stress, creating a need for irrigation. S_{SW} varies from day to day and depends on the crop type and TP:

S_{SW}=1-(DL_{ref}+0.04×(5-{{T_P}/0.95}))

DL_{ref} is a crop-type specific reference depletion level (essentially the fraction of AWC that can be depleted before stress occurs, defined downwards from wcfc). Allen et al. (1998) provide indicative values for DL_{ref} (cf. their Table 22). The S_{SW} equation is a slightly modified form of the original FAO-56 equation to account for the fact that only TP is used here. A typical S_{SW} profile is shown in Figure 3.1. By default, S_{SW} is limited to the range 0.2 – 0.9, but it can be further refined with the parameter P_{SSWCORR} (sswcorr in par.txt) to maximum 1.

If irrigation is needed, the required irrigation amount (W_{I,D,j}) can be calculated with three alternative methods in HYPE (chosen by the demandtype variable in MgmtData.txt):

(1) A constant W_{I,D,j} (defined by the irrdemand parameter in par.txt)

(2) Up to the field capacity: W_{I,D,j}=AWC-H

(3) Up to a defined fraction of S_{SW} (P_{iwdfrac}, iwdfrac in par.txt):

W_{I,D,j}=min[(S_{SW}×P_{SSWCORR}×AWC-H)×P_{iwdfrac},(AWC-H)]

The fraction can be larger than 1. For example, to irrigate to a level 10% above S_{SW}, P_{iwdfrac} =1.1. W_{I,D,j} is, however, limited to AWC.

Submerged crops

The irrigation of submerged crops aims to satisfy a target flooding level above the soil surface (Wisser et al., 2008). The target flooding level is a constant input parameter (P_{immdepth}, immdepth in par.txt). Irrigation is required if the water level of the top soil layer (H1) falls below the target flooding level:

if H_1  <WP_1+FC_1+EP_1+P_{immdepth} right irrigate

The irrigation water demand is equal to the amount needed to reach the target flooding level:

W_{I,D,j}=WP_1+FC_1+EP_1+P_{immdepth}-H_1

If the submerged season is shorter than the crop season, W_{I,D} during the non-submerged period is calculated in the same way as for non-submerged crops. To maintain a desired flooding level and simulate terracing, for example, it may be necessary to adjust the soil and runoff parameters of the class with the submerged crops. When W_{I,D,j} has been calculated for each irrigated class in the sub-basin, the total field-scale irrigation water demand for the sub-basin (W_{I,D}) is calculated:

W_{I,D}= sum{j=1}{N}{W_{I,D,j}}

Irrigation water withdrawal

Irrigation water can be abstracted from a set of water sources (Figure 2). Within a given sub-basin, water can be abstracted from the olake, the ilake, the main river, and from groundwater in a deep aquifer. In addition, water can be withdrawn from the olake and the main river of another sub-basin. These sources can be used on their own or in combination. Alternatively, HYPE can withdraw water from an unlimited source outside the model domain. This is specified with the irrunlimited code word in info.txt, and applies to all irrigated sub-basins.

Withdrawals are calculated just after the local discharge and the upstream discharge combine to flow into the main river of a given sub-basin.

Figure 2: Schematic illustration of the available irrigation sources in HYPE. Irrigation water can be withdrawn from: (i) dams in the sub-basin (olake and ilake), (ii) the main river in the sub-basin, (iii) groundwater in a deep aquifer, and (iv) dams (olake) and main rivers in other sub-basins.

Irrigation inefficiencies within the sub-basin

Before any withdrawal occurs, the field-scale W_{I,D} is scaled to sub-basin scale (W_{L,D}, the local sub-basin irrigation water demand). This is done in order to account for the often significant water losses between withdrawals and soil moisture replenishment. A simple user-provided scaling factor is applied:

W_{L,D}={W_{I,D}}/{E_L}

The local efficiency (E_L, local_eff in MgmtData.txt) represents the fraction of the withdrawn water within the sub-basin that infiltrates the irrigated soil. E_L accounts for losses in irrigation canals, ponds etc. within the sub-basin, and for on-field losses from irrigation equipment (e.g. sprinkler systems). W_{I,D} is not scaled if all withdrawals are from an unlimited source outside the model domain (i.e. E_L=1 if irrunlimited is y).

Withdrawal from sources within the sub-basin

The model first attempts to withdraw water from sources within the sub-basin where the demand originated. The user specifies the proportion of water to be withdrawn from surface water and deep aquifer groundwater sources, respectively, in the MgmtData.txt file (gw_part). A gw_part value between 0 and 1 represents the long-term average proportion of surface and deep aquifer groundwater withdrawals for irrigation within the sub-basin. A gw_part value of 0 indicates that all irrigation water exclusively comes from surface water sources, while a gw_part value of 1 indicates that all irrigation water exclusively comes from deep aquifers. Based on gw_part, W_{L,D} is split into the groundwater demand (W_{L,D,g}), and the surface water demand (W_{L,D,s}).

If any surface water demand exists, the model sequentially attempts to withdraw water from the olake, the ilake, and the main river. However, water withdrawal from olakes and ilakes is only calculated if the variable irrdam in MgmtData.txt is set to 1 for the sub-basin. Water withdrawals from the main river occur both from the inflow to the river reach and from the volume stored in the reach.

If the W_{L,D,s} volume is available at the source, the demanded water is withdrawn. Otherwise, only the available volume (V_s) is withdrawn. The withdrawal can also be scaled with the user-defined parameter P_{I,S} (pirrs in par.txt):

W_{L,A,s(1)}=min(W_{L,D,s},V_{s1})×P_{I,S}

W_{L,D,s2}=W_{L,D,s}-{W_{L,A,s(1)}}/{P_{I,S}}

where W_{L,A,s(1)} is the abstracted water from the first surface water source in the sub-basin, and W_{L,D,s2} is the residual surface water demand. The residual demands (W_{L,D,s2} and below W_{L,D,s,1}, W_{L,D,1} and W_{R,D2}) are calculated without the P_{I,S} scaling in order to prevent erroneous source compensation due to scaling. If any demand remains, the next surface water source is probed in the same manner:

W_{L,A,s(2)}=min(W_{L,D,s2},V_{s2} )×P_{I,S}

The total surface water withdrawal within the sub-basin (W_{L,A,s}), and the remaining surface water demand (W_{L,D,s,1}) is calculated accordingly, based on the abstractions from each source (k):

W_{L,A,s}= sum{k=1}{N}{W_{L,A,s(k)}}

W_{L,D,s,l}= W_{L,D,s}-{W_{L,A,s}}/{P_{I,S}}

If any groundwater demand exists, the model withdraws a user-specified fraction of W_{L,D,g} from an unlimited source outside the model domain or if an aquifer is simulated connected to the subbasin from this aquifer. The first case conceptually represents a large deep aquifer source, which is currently outside of the model domain. W_{L,a,g} is the abstracted groundwater and P_{I,G} (pirrg in par.txt) is the groundwater withdrawal fraction:

W_{L,A,g}=W_{L,D,g}×P_{I,G}

To simulate a more dynamic conjunctive use of groundwater and surface water sources, the model allows for compensation of remaining surface water demands from the groundwater source. This compensation is only allowed if both groundwater and surface water sources are used (0<gw_part<1), and if the irrcomp parameter is >0. The irrcomp parameter defines the degree of compensation allowed, i.e. the fraction of the residual surface water demand which can be met through source compensation. The compensation algorithm is as follows: if any surface water demand remains (W_{L,D,s,1} > 0) and the groundwater is not depleted, the groundwater withdrawal cycle is calculated once more using the scaled residual surface water demand. Finally, after possible source compensation, the remaining (surface) water demand at the sub-basin scale (W_{L,D,1}) is calculated.

Withdrawal from another sub-basin

The model can also simulate withdrawal from another sub-basin in the model domain (D_R, defined with the regsrcid input variable in MgmtData.txt). This withdrawal is calculated when the model reaches D_R in the calculation order, after possible local irrigation water abstractions in D_R and only if any irrigation demand remains in the sub-basin(s) connected to the regional source (W_{L,D,1} > 0). For each connected sub-basin (i), W_{L,D,l,i} is scaled to represent the regional-scale demand from that sub-basin (W_{R,D,i}). Again, a simple scaling factor is applied:

W_{R,D,i}={W_{L,D,l,i}}/{E_{R,i}}

The regional efficiency (E_{R,i}, region_eff in MgmtData.txt) represents the fraction of the withdrawn water at the regional source that reaches the connected sub-basin. E_{R,i} refers to the connected sub-basin. The regional scaling accounts for often significant water conveyance losses in large irrigation networks (in canals and dams etc.).

The total water demand from the regional source (W_{R,D}) is then calculated as the sum of the demand from each connected sub-basin, scaled by a parameter controlling the strength of the regional connection (P_regirr, regirr in par.txt):

W_{R,D}= (sum{i=1}{N}{W_{R,D,i}} )×P_{regirr}

The regional demand can be met from two sources in sub-basin D_R: the olake and the main river. If the regional source sub-basin has an olake, and if the irrdam input variable is set to 1 for that sub-basin, the model attempts to withdraw W_{R,D} first from the olake and then the residual from the main river. If not, the model only attempts to withdraw W_{R,D} from the main river. The regional abstraction (W_{R,a}) is limited by the volume available at the source (V_r) and the scaling parameter P_{I,S}:

W_{R,A1}=min(W_{R,D},V_{r1} )×P_{I,S}

W_{R,D2}=W_{R,D}-{W_{R,A1}}/{P_{I,S}}

W_{R,A2}=min(W_{R,D2},V_{r2} )×P_{I,S}

W_{R,A}=W_{R,A1}+ W_{R,A2}

where W_{R,A1} is the abstracted water from the first water source in D_R, V_{r1} the volume available at the first source, W_{R,D2} the residual regional water demand after withdrawal from the first source (but prior to the P_{I,S} scaling), W_{R,A2} the abstracted water from the second water source in D_R, and V_{r2} the volume available at the second source.

Substance concentrations of irrigation water withdrawals

The concentrations of the withdrawn water are the same as that of the irrigation water source. If the water originates from several sources, the volume-weighted concentration is calculated. If desired, the model can simulate sedimentation tanks, in which a defined fraction of the particulate phosphorous (pp) and organic nitrogen (on) settles:

C_{a,pp}=C_{src,pp}×(1-P_{cirrsink} )

C_{a,on}=C_{src,on}×(1-P_{cirrsink} )

where C_{,a} is the concentration of the abstracted water after settling, C_{src} is the concentration of the source, and P_{cirrsink} is the concentration reduction fraction (cirrsink parameter in par.txt). To use sedimentation tanks in a region, the concentration reduction fraction needs to be set in par.txt (0<cirrsink≤1).

Irrigation water application

In the calculation order, the irrigation water application occurs the next time step the model reaches the sub-basin from which the demand originated (typically the following day). The water is applied to the soils at the beginning of the soil balance calculations, before the calculation of the natural processes.

The regionally abstracted water (W_{R,A}) is first distributed to each connected sub-basin (i) according to their proportional demand, and then scaled to the local scale using the respective regional efficiency:

W_{R,A,i}=W_{R,A}× {W_{R,D,i}×P_{regirr}}/{W_{R,D}}×E_{R,i}

For a given sub-basin, the total amount of abstracted water available at the local scale (W_{L,A,i,tot}) is calculated and then scaled, using the local efficiency, to represent the water applied to the soil (W_{I,A,i}):

W_{L,A,i,tot}=W_{L,A,i}+W_{R,A,i}

W_{I,A,i}=W_{L,A,i,tot}×E_{L,i}

W_{I,A,i} is then distributed onto each irrigated class in proportion to its water demand:

W_{I,A,j}=W_{I,A,i}×{W_{I,D,j}}/{W_{I,D}}

W_{I,A,j} is added to the soil water of class j as additional infiltration. W_{I,A,j} is divided between the top two soil layers according to the epotdist function, beginning with the second layer. For unlimited irrigation: W_{I,A,j} =  W_{I,D,j} , P_{I,S} =1 and P_{I,G} = 1.

The water withdrawn from a regional source that does not reach connected sub-basins (W_{R,L}) evaporates at the regional source (D_R):

W_{R,L}=W_{R,A}-sum{i=1}{N}{W_{R,A,i}}

Similarly, water losses due to local inefficiencies (W_{L,L}) evaporate within the sub-basin itself. This applies to local losses from abstractions both within the sub-basin and from the regional source (where applicable):

W_{L,L}=W_{L,A,i,tot}-W_{I,A,i}

Evaporation due to regional and local inefficiencies proportionally concentrates substances in the withdrawn water. The substance concentrations in the irrigation water applications are hence higher than at the points of withdrawal (the mass remains the same while the volumes are reduced). However, if unlimited irrigation is simulated, the concentrations of the applied water are the same as in the layers to which water is added (i.e. causing no change in concentration).

Modules (file) Procedures Section
irrigation_module (irrigation.f90)calculate_irrigation irrigation water withdrawal
irrigation water application
apply_irrigationirrigation water application
irrigation_abstraction_sinksubstance concentrations of irrigation water withdrawals
calculate_irrigation_water_demandirrigation water demand
calculate_kcb
irrigation_season
immersion_season

References

Allen, R.G., L.S. Pereira, D. Raes, and M. Smith 1998. Crop Evapotranspiration (guidelines for computing crop water requirements), FAO Irrigation and Drainage Paper, No. 56, FAO, Rome, Italy, 300 pp.

Wisser, D., S. Frolking, E.M. Douglas, B.M. Fekete, C.J. Vörösmarty, and A.H. Schumann, 2008. Global irrigation water demand: Variability and uncertainties arising from agricultural and climate data sets, Geophysical Research Letters, Vol. 35, L24408, doi:10.1029/2008GL035296, 5 pp.

Point sources

Information on point sources is located in the file PointSourceData.txt.

Nutrient point sources

The model can handle up to three different types of point sources. They can used to simulate e.g. treatment plants, stormwater outlets, and industrial sources as separate types. All point sources have a constant flow, concentrations of total nitrogen and phosphorus, and fractions of IN and SP for a period of time. The time may be the whole simulation period, or different sources may be active during different parts of the simulation period. Point sources are added to the water in the main river.

Tracer T2 (water temperature) point sources

Water temperature may be added to the flow of nutrient point sources if T2 is simulated together with N and P. Water temperature point source may also be added on its own in the same way as nutrient point sources.

Tracer T1 point sources

Tracer T1 point source may be added in the same way as nutrient point sources. In addition, point sources of tracer T1 can be added to the local stream, the local lake, the main river or the outlet lake.

Negative point source

A point source with negative flow is treated as an abstraction of water. The abstraction can be made from the main river or the outlet lake. The water is removed from the source, while the concentration is kept.

Modules (file) Procedures
datamodule (data.f90)load_pointsourcedata
npc_surfacewater_processes (npc_sw_proc.f90)add_point_sources_to_main_river
surfacewater_processes (sw_proc.f90)point_abstraction_from_main_river
point_abstraction_from_outlet_lake
tracer_processes (t_proc.f90)add_tracer_point_source_to_river
add_tracer_point_source_to_lake
start/hype_model_description/hype_human_water.txt · Last modified: 2017/05/31 12:00 by jandersson