CN109616223A - A kind of improvement and application method based on SWAT modeling River Basin Hydrology process - Google Patents

Abstract

The present invention passes through the key parameter that land use/cover is that hydrologic process is simulated, the hydrologic processes such as influence retains, infiltrates, evapotranspiration and the recharge of ground water, strong region is being interfered by mankind's activity, acute variation can occur for land use/cover type, and then influence watershed hydrologic cycle process.SWAT is as a widely applied hydrological distribution model, within its simulation phase, land use/cover data cannot be called year by year, have ignored the temporal variation of land use/cover, it may be will affect in the application of land use/cover variation violent area, herein for this disadvantage of SWAT model, it is carried out to improve and develop to call land use/cover data LU-SWAT model year by year.Change violent Black River in land use/cover to be compared the Runoff Simulation effect of SWAT and LU-SWAT model, discovery LU-SWAT model is more suitable for Black River water circulation faults.

Description

A kind of improvement and application method based on SWAT modeling River Basin Hydrology process

Technical field

The present invention relates to SWAT model applied technical fields, more particularly to one kind to be based on SWAT modeling River Basin Hydrology mistake The improvement of journey and application method.

Background technique

There is distribution (or semi) hydrological model of physical basis that can clearly reflect the sky of River Basin Hydrology process Between variability, have in terms of the variation of River Basin Hydrology process under the influence ofs analyzing and researching climate change, mankind's activity etc. important Using.SWAT (Soil and Water Assessment Tools) model as a typical hydrological distribution model, by The features such as simulation effect open, powerful in its code, in many basins is satisfactory, obtains in worldwide Extensive utilization.Such as Liu Changming is in Yellow River source with SWAT modeling its moon, year channel flow amount；Jin etc. is in Heihe Upstream is using SWAT model analysis influence of the soil attribute special heterogeneity to hydrologic process；Wang Sheng etc. utilizes SWAT model point Influence of the climate change to Basin of Huaihe River middle and upper reaches water balance factors is analysed.Land use/cover variation is that global environment becomes The important component and main cause of change.A large number of studies show that land use/cover is the key that hydrologic process simulation ginseng Number is influenced to retain, be infiltrated, the hydrologic processes such as evapotranspiration and the recharge of ground water.SWAT as a distributed hydrological model, it Entire research basin is divided into several sub-basins according to features such as terrain factor, the network of waterways distributions in basin first.It is basic herein On, further Hydrologic response units are divided by the land use/cover type, soil types and gradient area threshold in basin (Hydrological Response Units,HRUs).For the terseness that model calculates, within the simulation phase of SWAT model, Assuming that land use/cover the type, soil types and the gradient in basin are constant.However, being interfered by force certain by mankind's activity Strong basin, acute variation occurs for land use/cover type, and then influences watershed hydrologic cycle process.And SWAT model is just neglected This point is omited: no matter how long it simulates the phase, using only the land use/cover data in a certain year.This makes model in soil Using/cover variation acutely area apply existing defects.Currently, part research considered in modeling land use/ The variation of cover: Zhou etc. passes through SWAT modeling in Process of Urbanization in Yangtze River in China Delta Area, basin water The variation of literary process；Li and Zhou has studied basin soil under the effect of human activity in the Yanhe River basin of China, by SWAT model The variation of ground utilization/cover and its influence of watershed water balance；It opens icepro etc. and simulates soil under multiple scenes in SWAT model Utilize the influence after changing to the Middle Reaches of Heihe River hydrologic process；Wang Xue etc. is based on SWAT model and has studied Baima river valley The influence of land use/cover change on runoff.In these researchs, due to research area soil in the SWAT model hypothesis simulation period Ground utilization/cover type is constant, therefore the simulation period has been carried out uniform division.Such as: if simulation the period be nineteen ninety-five extremely 2014, this is first divided into 5 equal sub-periods: 1995-1999,2000-2004,2005-2009,2010-2014 for 20 years, Land use/cover the data for being utilized respectively a certain year in this 5 simulation periods again respectively drive SWAT mould within these periods Type.This mode has paid close attention to variation of the land use/cover within the entirely simulation period substantially, but model manipulation person is come It says, this com-plicates the foundation of hydrological model and dry run process.And this mode may still be difficult to reflect soil benefit With/cover acute variation.

Therefore, the characteristics of this research cannot reflect land use in the simulation period/cover variation for SWAT model, HRUs generating mode, model calculating etc. have carried out corresponding improvement to SWAT model, make it can be year by year within the simulation period Call land use/cover data.Finally, improved SWAT model is applied to land use/cover class in recent decades Type, which has occurred the complicated and Black River of acute variation and verifies its, simulates effect.

Summary of the invention

In order to achieve the above object, the present invention provides a kind of improvement and application based on SWAT modeling River Basin Hydrology process Method.

A kind of improvement and application method based on SWAT modeling River Basin Hydrology process,

Hydrological model of the SWAT as a semi, first by entire research basin according to the landform in basin because Son, network of waterways distribution characteristics are divided into several sub-basins, and on this basis, SWAT further presses the land use pattern in basin, soil Earth type and gradient area threshold divide Hydrologic response units, i.e. HRUs.Since SWAT assumes land use class in the simulation period Type, soil types and the gradient are constant, and the HRUs of generation will not change attribute because of the difference for simulating the time, each individual On HRU, SWAT is using conceptual model estimation rainfall, production stream, silt etc..These carry out concentration of channel calculation after calculating completion, Finally acquire the flow, quantity of sand and mud and pollution loading amount of Outlet Section；

The River Basin Hydrology process of SWAT modeling can be divided into:

1) the land face part of water circulation produces stream and slope concentration part, it controls main stem in each sub-basin The input quantity of water, sand, nutriment etc.；

2) it is defeated to determine that the substances such as water, sand are exported from the network of waterways to basin for the water surface part of water circulation, i.e. concentration of channel part It moves；

According to principle of water balance, the water calculating of SWAT model follows following formula: Tu=T0+ ∑ (S eay-R tvsg- Eu-Ttffp-R w) ui=1 (1)；

SWAT model refinement is divided into following components:

Part one: quantity, area and the spatial position for the HRUs that utility many years land use/cover data generate are not Become；

Part two: before generating HRUs, the land use data year by year in the period will be studied first and carries out space overlapping (successive according to the time) generates superimposed land-use map and obtains its corresponding attribute list, have patch number in the attribute list And the attributes such as land use pattern of different year corresponding to each patch.By the land-use map and soil after space overlapping Earth type map, slope map carry out space overlapping again, each spot in the attribute list for calling back to generate in additive process year by year The land use pattern of the corresponding different year of block, generates the HURs of different year；

Part three；Need to modify the operation process of original SWAT model, original SWAT model first carries out before the computation Parameter initialization enters recycle day afterwards, and compared with original SWAT model, new SWAT model is added before parameter initialization One layer of loop structure, makes program read the input data (HRUs data) of current year before initiation parameter, then counts It adds into a day cycle stage；If the one's last year of the simulation period upper one year in program read time, then terminate to calculate.Otherwise Model simulation results are written file after terminating and read again the input data in this time simultaneously by the simulation of annual last day Corresponding parameter initialization is carried out, day circulation is entered back into；Structural modification SWAT source code is calculated according to new model, it can be true The positive variation year by year for embodying land use data.

As further explanation, the HRUs is that entire research basin is special according to the terrain factor in basin, network of waterways distribution Sign is divided into several sub-basins, and on this basis, SWAT further presses the land use pattern, soil types and gradient face in basin Product threshold value divides Hydrologic response units.

As further explanation, St is soil moisture content/mm in the formula (1)；S0 indicate early period soil moisture content/ mm；The time step of t expression modeling；Rday indicates i-th day precipitation/mm；Qsurf indicates i-th day rainwash Amount/mm；Et indicates actual evaporation capacity/mm；Sseep indicates i-th day soil permeability amount/mm；Qgw indicates Dryweather flow/mm.

As further explanation, environment Visual is developed in windows platform application with Fortran language It modifies modification and after compiling successfully to SWAT source code (Rev.635) according to the above structure in Studio 2010, replacement is former The SWAT.exe executable program of beginning.Since improved SWAT model realization is to land use/cover data tune year by year With.

Beneficial effects of the present invention:

1. the characteristics of research cannot reflect land use in the simulation period/cover variation for SWAT model, in HRUs Generating mode, model calculating etc. have carried out corresponding improvement to SWAT model, realize it within the simulation period to soil benefit With the calling year by year of/cover data, LU-SWAT model is developed；

2. SWAT model and LU-SWAT model are applied to land use/cover variation complexity, acutely and by agricultural irrigation It influences big Black River and compares the simulation effect of two models, discovery can call land use/cover data year by year LU-SWAT model is more suitable for the river region of Black River and similar hydrologic condition.

Detailed description of the invention

Fig. 1 is improved SWAT model schematic diagram of calculation flow；

Fig. 2 is Irrigation Area in Middle Reaches of Heihe River irrigation volume icon；

Fig. 3 is that HRUs generating process modifies schematic diagram；

Fig. 4 is Black River dynamic degree of land use (%) chart；

Fig. 5 is SWAT parameters sensitivity analysis result chart；

Fig. 6 is that the monthly runoff of Black River justice gorge station SWAT and LU-SWAT modeling and measured value comparison are illustrated Figure；

Fig. 7 is that the annual runoff of Black River justice gorge station SWAT modeling and measured value compare；

Fig. 8 is Black River justice gorge station SWAT and LU-SWAT the model moon, annual runoff simulation effect assessment.

Specific embodiment

Technical solution of the present invention work specifically, is completely illustrated with reference to the accompanying drawing.

Black River is chosen herein as case study on implementation.

SWAT model brief introduction:

Hydrological model of the SWAT as a semi, it first by entire research basin according to the landform in basin because The features such as son, network of waterways distribution are divided into several sub-basins.On this basis, SWAT further press basin land use pattern, Soil types and gradient area threshold divide Hydrologic response units, i.e. HRUs.Since SWAT assumes land use in the simulation period Type, soil types and the gradient are constant, therefore the HRUs generated will not change attribute because of the difference for simulating the time, each On individual HRU, SWAT is using conceptual model estimation rainfall, production stream, silt etc..These carry out the concentration of channel after calculating completion Calculation.Finally acquire the flow, quantity of sand and mud and pollution loading amount of Outlet Section.

The River Basin Hydrology process of SWAT modeling can be divided into: 1. the land face part of water circulation (produces stream and slope concentration portion Point), it controls the input quantity of the water of main stem, sand, nutriment etc. in each sub-basin；2. the water surface part of water circulation (concentration of channel part), it determines the defeated shifting [9] that the substances such as water, sand are exported from the network of waterways to basin.

According to principle of water balance, the water calculating of SWAT model follows following formula:

St=S0+ ∑ i=1t (Rday-Qsurf-Et-Sseep-Qgw) (1)

In formula: St is soil moisture content/mm；S0 indicates soil moisture content/mm of early period；The time of t expression modeling Step-length；Rday indicates i-th day precipitation/mm；Qsurf indicates i-th day flow path surface/mm；Et indicates actual evaporation Amount/mm；Sseep indicates i-th day soil permeability amount/mm；Qgw indicates Dryweather flow/mm.Such as Fig. 2.

In addition, the irrigation operation in SWAT model has man-made irrigation operation and automatic irrigation to operate 2 kinds, people is taken in this research Work irrigates the mode of operation.SWAT model arranges man-made irrigation to operate by date or heat unit, the irrigation of user's input Amount refers to the water for reaching soil, and SWAT model considers the loss water from water source to soil by irrigation efficiency, including passes Defeated loss and evaporation loss.The score of duty is accounted for as rainwash ratio using the water that rainwash leaves field, it is remaining Moisture penetrate into field soil, calculated with the soil water calculation method of SWAT model.In this way soil slope surface can be reached with real surface Agriculture flood irrigation mode.

SWAT model refinement:

The basic computational ele- ment of SWAT model is HRUs.Each HRU is jointly fixed by land use, soil types and the gradient Justice.The same HRU is made of multiple grids, these grids can be spatially adjacent, can also be separated from each other.The number of HRUs Amount, area and spatial position are by patch quantity, plaque area and the patch position in land-use map, soil type map and slope map Set co-determination.In data of this research for generating HRUs, soil types and the gradient are constant, and land use changes year by year. So that SWAT model is called land use/cover data year by year, first has to what utility many years land use/cover data generated The quantity of HRUs, area and spatial position are constant.

As shown in figure 3, for quantity, area and the sky of utility many years land use/cover data generation HRUs Between position it is constant, before generating HRUs, first by study the period in land use data year by year carry out space overlapping (according to Time is successive), it generates superimposed land-use map and obtains its corresponding attribute list, there is patch number and each in the attribute list The attributes such as the land use pattern of different year corresponding to a patch.By the land-use map and soil types after space overlapping Figure, slope map carry out space overlapping again, and each patch is corresponding in the attribute list for calling back to generate in additive process year by year Different year land use pattern, generate the HURs of different year.So, by the land use of script different year Figure corresponds on superimposed figure, the land-use map in each time patch quantity, area and spatial position having the same, but It is that the corresponding attribute of each patch (land use pattern) may be different due to the time.It ensure that each year ultimately produced The quantity of HRUs, area and spatial position are constant, and providing to the subsequent calculating of the SWAT model based on land use change survey can Energy.

Realize that SWAT model to land use/cover data calling year by year, will not only redefine and generate HRUs's Process so that keep quantity, area consistent with spatial position with the HRUs that different year land use data generates.It is more important , need to modify the operation process of original SWAT model.Original SWAT model first carries out parameter initialization before the computation, after It is recycled into day.This can not achieve the calling year by year of land use.

Fig. 1 is the operation process for having made corresponding improved SWAT model.Compared with original SWAT model, new SWAT Model joined one layer of loop structure before parameter initialization, and program is made to read the defeated of current year before initiation parameter Enter data (HRUs data), then calculates and enter cycle stage day.If the simulation period upper one year in program read time One's last year then terminates to calculate.Otherwise file and again is written into model simulation results after the simulation of annual last day terminates The secondary input data for reading this time simultaneously carries out corresponding parameter initialization, enters back into day circulation.Knot is calculated according to new model Structure modifies SWAT source code, can really embody the variation year by year of land use data.Existed in this research with Fortran language Windows platform application is developed in environment Visual Studio 2010 according to the above structure to SWAT source code (Rev.635) it modifies.The subprograms such as main modification simulate.f, main.f.After modifying and compiling successfully, replace original SWAT.exe executable program.Since improved SWAT model realization is to land use/cover data calling year by year, Therefore improved SWAT model is named as LU-SWAT herein.

The modeling effect assessment factor:

Choose that assorted coefficient (Nash-Sutcliffe Efficiency, NSE), error percentage (Percent Bias, PBIAS), root-mean-square error and standard error ratio (Ratio of the Root Mean Square Error to the Standard Deviation of Measured Data, RSR) evaluation SWAT model applicability.The value range of NSE is (- ∞, 1), NSE closer to 1, then indicate that the simulation effect of SWAT model is good and model it is with a high credibility；NSE is closer to 0.5, then Model simulation results are indicated close to the average level of observation, model total result is credible.And PBIAS is between -10% to 10% Between, then show that modeling effect is good.RSR is smaller, shows that modeling effect is better.For the detailed of 3 evaluation points Calculating process and meaning (with reference to the research of Moriasi etc.).

Interpretation of result

Black River land use/cover variation:

Fig. 4 shows Black River dynamic degree of land use, and there are following rules for the land use change survey of Black River: ploughing Ground, forest land and cities and towns are increase trend, but the ratio that cities and towns area accounts for entire research area's area is smaller.1990-2009 The land use change survey of Black River is mainly manifested in: arable land, forest land and cities and towns are increase trend, wherein, cities and towns area increases Add comparatively fast, amplification is larger, increases more than 1 times of original area.But the ratio that cities and towns area accounts for research area's gross area is smaller.Forest land face Product increases 55.70% in 20 years in increase trend is stablized.Arable land face increases 14.78%.Meadow and bare area area are to reduce Gesture, wherein grassland area reduces 10.15%, and bare area area reduces 21.11%.Water surface area increases 14.10%.Herein Described is each land use/20 years general morphologictrends of cover type, but as seen from Table 2, and any land use/ The variation year by year of cover type is different.Such as cultivated area increases year by year, but increased amplitude varies；Meadow face Product is generally reduction trend, but there are increase trend in individual times.Therefore, more illustrate the land use/cover data in a certain year without Method represents regional land use/cover situation of long duration.It should be noted that land use/cover variation itself can be right Water cycle process has an impact, and because basin caused by the variation of land use/cover takes water use variation also shadow to a certain extent Xiangshui County's circulation.Such as in Black River, cultivated area increases 14.78% from 1990-2009, this can further influence agriculture Industry irrigation volume.And in Black River, agricultural irrigation is [16,18] an important factor for influencing watershed hydrologic cycle.In addition, SWAT model It is 2661 in the HRUs quantity that Black River divides before improvement, and passes through improved LU-SWAT model, HRUs number It is 4816 that amount, which increases,.In the case where soil types and constant terrain data, illustrate there are 2155 in LU-SWAT model Land use/cover on HRUs is changed between 20 years.This has inevitable influence to model calculating.

Simulation Contrast on effect of the SWAT and LU-SWAT in Black River

Model tune ginseng

Since LU-SWAT model is compared with SWAT model, there is no the change of physical process, calculation method etc., because This this research thinks that SWAT is consistent with the responsive parameter of LU-SWAT.Herein with reference to the research of White and Chaubey, choose 22 parameters relevant to water cycle process have carried out sensitivity analysis [20] in SWAT model.Sensitivity Analysis is using more First regression model returns the parameter that Latin hypercube (Latin Hypercube) sampling generates with target function value:

G=δ+∑ i=1n θ iai (2)

In formula: g is objective function；δ, θ i are the coefficient of regression equation；Ai is parameter value；N is number of parameters.T value in table 3 For detecting the sensibility of each parameter, absolute value shows that more greatly parameter is more sensitive, and p value is for determining susceptibility in table 3 Significance degree, value show more significant closer to 0.

Fig. 5 shows parameters sensitivity analysis result.For SWAT model, this research is chosen most sensitive preceding 10 in Fig. 5 A parameter carries out calibration, i.e. effective infiltration coefficient (CH_K2) of main stem alluvial deposits, initial SCS runoff when moisture condition II Curve number (CN2), the base flow factor (ALPHA_BF), the Manning coefficient (CH_N2) of main stem occur shallow-layer needed for returning stream and contain The watermark threshold (GWQMN) of water layer, the snow melt factor (SMFMN) on annual December 21, the re-evaporation coefficient (GW_ of underground water ), REVAP the time delay (GW_DELAY) of underground water, snowfall temperature (SFTMP) and the accumulated snow temperature lag factor (TIMP).Water The parameter calibration process of literary model is not the fitting simply to measured data, but comprehensively considers drainage characteristics, is not being surpassed Parameter calibration is carried out under the premise of Rational Parameters range out, makes analogue data close to measured data.Therefore, carrying out parameter herein Value range restriction is carried out to each responsive parameter according to Heihe River basin existing research first when calibration, then utilizes Arnold The manual parameter distribution method of the multi-step mentioned in research Deng [9] carries out.

Since LU-SWAT model calls land use/cover data year by year, therefore in parameter calibration, in order to embody soil Ground utilization/cover changes the influence to model parameter, this research makes different soil benefits using the method for " dynamic state of parameters calibration " Different best parameter groups is corresponded to data.As described above, 10 sensitivities are had chosen when carrying out parameter calibration to SWAT Parameter.Therefore, when carrying out parameter calibration to LU-SWAT model, choose 4 influenced in 10 parameters by land use/cover A parameter carries out calibration (CH_K2, CN2, ALPHA_BF, GW_REVAP), and other parameters value is constant.With simulation obtain 20 years by Year land use/cover data respectively drive LU-SWAT model, and survey mountainous runoff Data correction using just gorge hydrometric station LU-SWAT model obtains 20 groups of best parameter groups.

SWAT and LU-SWAT simulates Contrast on effect

Fig. 6 and Fig. 8 shows SWAT model to the simulation effect of Black River justice gorge hydrometric station monthly runoff.By table 4 Known to, NSE value is 0.67；PBIAS value reaches 4.33%；RSR value reaches 0.57.Therefore, SWAT model goes out the Black River moon The simulation effect of path flow is preferable.The SWAT modeling effect of Black River justice gorge station annual runoff also preferably (Fig. 6, Table 4): NSE value is that 0.72, PBIAS value reaches 4.33%, RSR 0.53.Compared with Black River existing research [12,22], this The simulation effect of SWAT model is slightly good in research.Preceding to have addressed, in Black River, agricultural irrigation is the weight for influencing watershed hydrologic cycle Want factor.This research has input detailed agricultural irrigation data in SWAT model foundation time-division difference irrigated area, considers agriculture filling Variation spatially is irrigate, therefore the SWAT model that this research is established obtains better Runoff Simulation effect in Black River.

Fig. 6,8 and table 7 also show the Black River justice gorge station LU-SWAT model moon, annual runoff simulation effect.With SWAT model is compared, and on moon scale, it be 7.42%, RSR value is 0.51 that NSE value, which has reached 0.74, PBIAS value,；On year scale, It is 7.42%, RSR value is 0.48 that NSE value, which has reached 0.77, PBIAS value,.That is, LU-SWAT model is in Black River Runoff Simulation effect is more preferable.This may be because the run-off in Black River justice gorge is mainly by the shadow of the basin agricultural irrigation It rings, and agricultural irrigation water is mainly influenced by current year farmland area, crop species etc..In this research, not only make LU-SWAT mould Type has invoked land use/cover data year by year, also when calling the HRUs in corresponding time for different irrigated areas, Different Crop by Year has input detailed agricultural irrigation data.Original SWAT model is also difficult on reflection HRU since its HRUs is remained unchanged The change year by year for the irrigation volume that may occur.But improved LU-SWAT model not only can reflect basin land use/cover Variation, can also embody the agricultural irrigation of Black River difference HRUs situation of change year by year.This enable LU-SWAT model more Accurately simulate the changes in runoff of Black River.Therefore, compared with SWAT model, land use/cover can be called year by year The LU-SWAT model of data changes violent Black River more suitable for land use/cover.

It is the characteristics of this research cannot reflect land use in the simulation period/cover variation for SWAT model, raw in HRUs At mode, model calculate etc. corresponding improvement has been carried out to SWAT model, realize its simulation the period in land use/ The calling year by year of cover data, develops LU-SWAT model；

SWAT model and LU-SWAT model are applied to land use/cover variation complexity, acutely and by agricultural irrigation shadow It rings big Black River and compares the simulation effect of two models, discovery can call land use/cover data LU- year by year SWAT model is more suitable for the river region of Black River and similar hydrologic condition.

Obviously, described embodiment is only a part of the embodiments of the present invention, rather than whole embodiments.Based on this Embodiment in invention, all other reality obtained by those of ordinary skill in the art without making creative efforts It applies, belongs to protection scope of the present invention.

Claims (4)

1. a kind of improvement and application method based on SWAT modeling River Basin Hydrology process, which is characterized in that

Hydrological model of the SWAT as a semi, first by entire research basin according to the terrain factor in basin, Network of waterways distribution characteristics is divided into several sub-basins, and on this basis, SWAT further presses the land use pattern in basin, soil class Type and gradient area threshold divide Hydrologic response units, i.e. HRUs, since SWAT assumes land use pattern, soil in the simulation period Earth type and the gradient are constant, and the HRUs of generation will not change attribute because of the difference for simulating the time, on each individually HRU, SWAT is using conceptual model estimation rainfall, production stream, silt etc..These carry out concentration of channel calculation after calculating completion, finally ask Obtain flow, quantity of sand and mud and the pollution loading amount of Outlet Section；

The River Basin Hydrology process of SWAT modeling can be divided into:

1) the land face part of water circulation, that is, produce stream and slope concentration part, it control the water of main stem in each sub-basin, The input quantity of sand, nutriment etc.；

2) water surface part of water circulation, i.e. concentration of channel part determine the defeated shifting that the substances such as water, sand are exported from the network of waterways to basin；

According to principle of water balance, the water calculating of SWAT model follows following formula: Tu=T0+ ∑ (Seay-Rtvsg-Eu- Ttffp-Rw) (1) ui=1；

SWAT model refinement is divided into following components:

Part one: the quantity of the HRUs that utility many years land use/cover data generates, area and spatial position are constant；

Part two: before generating HRUs, the land use data year by year in the period will be studied first and carries out space overlapping, is generated Superimposed land-use map simultaneously obtains its corresponding attribute list, has corresponding to patch number and each patch in the attribute list The attributes such as the land use pattern of different year.Again with soil type map, slope map by the land-use map after space overlapping Space overlapping is carried out, the soil of the corresponding different year of each patch in the attribute list for calling in additive process back to generate year by year Ground use pattern generates the HURs of different year；

Part three: needing to modify the operation process of original SWAT model, and original SWAT model first carries out parameter before the computation Initialization enters recycle day afterwards, and compared with original SWAT model, new SWAT model joined one before parameter initialization Layer loop structure, so that program is read the input data (HRUs data) of current year before initiation parameter, then calculate into Enter a day cycle stage；If the one's last year of the simulation period upper one year in program read time, then terminate to calculate；Otherwise annual Model simulation results are written file after terminating and read again input data and the progress in this time by the simulation of last day Corresponding parameter initialization enters back into day circulation；Structural modification SWAT source code is calculated according to new model, it can real body The variation year by year of existing land use data.

2. a kind of improvement and application method based on SWAT modeling River Basin Hydrology process according to claim 1, special Sign is:

The HRUs is entirely to study basin to be divided into several sub-basins according to the terrain factor in basin, network of waterways distribution characteristics, On the basis of this, it is single that the land use pattern, soil types and gradient area threshold that SWAT further presses basin divide hydrology response Member.

3. a kind of improvement and application method based on SWAT modeling River Basin Hydrology process according to claim 1, special Sign is:

St is soil moisture content/mm in the formula (1)；S0 indicates soil moisture content/mm of early period；T indicate modeling when Between step-length；Rday indicates i-th day precipitation/mm；Qsurf indicates i-th day flow path surface/mm；Et indicates actual steaming Hair amount/mm；Sseep indicates i-th day soil permeability amount/mm；Qgw indicates Dryweather flow/mm.

4. a kind of improvement and application method based on SWAT modeling River Basin Hydrology process according to claim 1, special Sign is:

With Fortran language according to the above knot in windows platform application exploitation environment Visual Studio 2010 Structure modifies modification and after compiling successfully to SWAT source code, replaces original SWAT.exe executable program, improved SWAT model realization is to land use/cover data calling year by year.