CN102103077A - MODIS data-based agricultural drought monitoring method - Google Patents

Abstract

The invention discloses a moderate-resolution imaging spectroradiometer (MODIS) data-based agricultural drought monitoring method. In the farmland drought monitoring process, an agricultural drought index is determined by a crop water supply index and a rainfall anomaly index; in the drought monitoring process, a vegetation index and the surface temperature are inverted by MODIS data, and the crop water supply index is calculated by the vegetation index and the surface temperature; the rainfall anomaly index is calculated by rainfall data; and drought severity is determined by dividing the level of the drought index.

Description

A kind of agricultural arid monitoring method based on the MODIS data

Technical field

The present invention relates to a kind of agricultural drought monitoring method.

Background technology

The generating process of arid is potential, is not easy to find; The occurrence characteristics of agricultural drought is that coverage is big, brings serious catastrophic effect and economic loss; Research, estimate that arid takes place and the process of development, can take the corresponding drought resisting measure of preventing and reducing natural disasters, reduce the agricultural disaster loss.Along with the development of remote sensing technology, remote sensing is with its dynamic, real-time, multispectral, cheap advantage, for new approach has been opened up in the damage caused by a drought monitoring.Vegetation index that remote sensing is obtained and surface temperature are to describe two ten minutes important parameters of earth surface feature, when therefore arid takes place, can disclose the plant physiology off-note by the variation of vegetation index or surface temperature, reflect farmland hydro-thermal stress state indirectly.Can be divided into according to the employed wave band of remotely-sensed data: visible light, near infrared, thermal infrared, microwave etc.Because employed wave band difference has produced numerous models and method.As water deficit exponential model, temperature vegetation drought index model etc.Though it is a lot of to be used for the model of agricultural drought disaster monitoring, great majority are experimental research.Therefore it is very necessary to explore a kind of accurate relatively agricultural drought monitoring method.

Summary of the invention

The purpose of this invention is to provide a kind of new method of utilizing remote sensing technology to carry out face of land draught monitor, this agricultural drought monitoring utilizes crop water supply exponential sum rainfall anomaly index to determine agriculture damage caused by a drought index.When carrying out the damage caused by a drought monitoring, utilize MODIS data inversion vegetation index and surface temperature, utilize vegetation index and surface temperature to calculate crop water supply index.Utilize precipitation data computation precipitation anomaly index.At last the damage caused by a drought index is carried out rank and divide to determine the arid order of severity.。

The present invention adopts and monitors based on the agricultural arid monitoring model of vegetation index and surface temperature, in the process of monitoring the model Ecological Parameter is improved, the relevant parameter of the face of land damage caused by a drought of explaining is determined as follows: based on the spectral reflectance data of each wave band pixel of the pairing MODIS of monitored farmland massif;

A kind of agricultural arid monitoring method based on the MODIS data mainly is to inquire into a kind of objective, dynamic, real-time, accurate, monitoring method of being easy to realize by the remote sensing technology means on big regional scale.Adopt in the method based on the agricultural arid monitoring model of vegetation index and surface temperature and monitor, in the process of monitoring the model Ecological Parameter is improved, the relevant parameter of the face of land damage caused by a drought of explaining is determined as follows:

Spectral reflectance data based on each wave band pixel of the pairing MODIS of monitored farmland massif;

Obtain EVI, T _s, VSWI and SDI parameter, remotely-sensed data to selected wave band is carried out unified pre-service, mainly comprise according to a conventional method and carry out (in remote sensing softwares such as ERDAS or ENVI, handling) radiant quantity calculating, geometry correction, cloud detection, projection conversion, so that further mate calculating; When calculating rainfall anomaly index, at first the rainfall data should be carried out interpolation processing, obtain precipitation anomaly figure; The unified raster data that changes into after interpolation is finished, so as with the remotely-sensed data comparative analysis.

1) calculating of vegetation index:

$\mathrm{EVI}=G\×\frac{{\mathrm{\ρ}}_{\mathrm{NIR}}-{\mathrm{\ρ}}_{\mathrm{Red}}}{{\mathrm{\ρ}}_{\mathrm{NIR}}+C_1\×{\mathrm{\ρ}}_{\mathrm{Red}}-C_2\×{\mathrm{\ρ}}_{\mathrm{Blue}}+L}$

EVI is for strengthening vegetation index, ρ _NIR, ρ _RedAnd ρ _BlueBe respectively the spectral reflectivity of near-infrared band, red spectral band and the blue wave band pixel of the pairing MODIS of monitored farmland massif; L is that background is adjusted item; C ₁And C ₂Be fitting coefficient; G is a gain factor; When calculating MODIS-EVI, L=1, C ₁=6, C ₂=7.5, G=2.5; Calculate the EVI data of pairing each the pixel point of remote sensing image of monitored farmland massif respectively;

2) Surface Temperature Retrieval

Surface temperature is a basic parameter of agricultural drought disaster monitoring, its computing method adopt (Qin Zhihao such as Qin Zhi person of outstanding talent, Gao Maofang, Qin Xiaomin etc., surface temperature remote sensing inversion method in the agricultural drought disaster monitoring-with MODIS data instance [J]. the disaster journal, 2005,14 (4): the 64-71.) window algorithm of splitting based on middle large scale MODIS data of Ti Chuing, its computing formula is:

T _s＝A ₀+A ₁T ₃₁-A ₂T ₃₂

A ₀＝E ₁a ₃₁-E ₂a ₃₂

A ₁＝1+A+E ₁b ₃₁

A ₂＝A+E ₂b ₃₂

A＝D ₃₁/E ₀

E ₁＝D ₃₂(1-C ₃₁-D ₃₁)/E ₀

E ₂＝D ₃₁(1-C ₃₂-D ₃₂)/E ₀

E ₀＝D ₃₂C ₃₁-D ₃₁C ₃₂

C _t＝ε _iτ _i

D _t＝[1+(1-ε _i)τ _i]

T in the formula _sBe surface temperature (K), T ₃₁And T ₃₂Be respectively the brightness temperature of MODIS the 31st and 32 wave bands, A ₀, A ₁And A ₂Be the parameter of splitting window algorithm, a ₃₁, b ₃₁, a ₃₁And b ₃₂Be constant, the desirable a of difference in surface temperature 0-50 ℃ scope ₃₁=-64.60363, b ₃₁=0.440817, a ₃₂=-68.72575, b ₃₂=0.473453; The calculating of other intermediate parameters sees the division window algorithm (Qin Zhihao at the MODIS data such as Qin Zhihao for details, Gao Maofang, Qin Xiaomin etc., the surface temperature remote sensing inversion method in the agricultural drought disaster monitoring-with MODIS data instance [J]. the disaster journal, 2005,14 (4): 64-71.); Utilize formula to calculate the T of pairing each the pixel point of remote sensing image of monitored farmland massif respectively then _sData;

Wherein i is meant the 31st and 32 wave bands of the pairing MODIS image of monitored farmland massif, is respectively i=31 or 32; τ _iBe the atmospheric transmittance of the wave band i of pairing each the pixel point of remote sensing image of monitored farmland massif, ε _iIt is the face of land emissivity of the wave band i of pairing each the pixel point of remote sensing image of monitored farmland massif.

ε _t＝ε _iw+P _vR _vε _iv+(1-P _v)R _sε _is

ε _Iw, ε _IvAnd ε _IsBe respectively water body, vegetation and the exposed soil of pairing each the pixel point of remote sensing image of monitored farmland massif face of land emissivity, get ε respectively at wave band i wave band _31w=0.99683, ε _32w=0.99254, ε _31v=0.98672, ε _32v=0.98990, ε _31s=0.96767, ε _31s=0.97790; R _vAnd R _sBe respectively the vegetation of the pairing remote sensing image of monitored farmland massif and the radiation ratio of exposed soil, (Qin Z such as Qin Zhihao, Karnieli A.Progress in the Remote Sensing of Land SurfaceTemperature and Ground Emissivity Using NOAA-AVHRR[J] .InternationalJournal of Remote Sensing, 1999,20 (12): 2367-2397.), the R of calculating _v=0.99240, R _s=1.00744.P _vBe the vegetation coverage of pairing each the pixel point of monitored farmland massif, estimate by vegetation index:

$P_v=\frac{\mathrm{NDVI}-{\mathrm{NDVI}}_s}{{\mathrm{NDVI}}_v-{\mathrm{NDVI}}_s}$

NDVI is the vegetation index of pairing each the pixel point of monitored farmland massif in the formula, NDVI _vAnd NDVI _sBe respectively the NDVI value of dense vegetation covering and complete exposed soil pixel, get NDVI usually _v=0.674, NDVI _s=0.039 (Sun Rui, Zhu Qijiang. the analysis [J] of the clean primary productivity of clean primary productivity model of vegetation and China. Beijing Normal University's journal (natural science edition), 1998,34:132-137.);

$\mathrm{NDVI}=\frac{B_2-B_1}{B_2+B_1}$

B in the formula ₁And B ₂It is respectively the reflectivity of MODIS image the 1st and 2 wave bands.

Present near linear relationship between atmospheric transmittance and the atmosphere vapour content, according to Qin Zhihao etc. (

Qin?Z，Karnieli?A，Berliner?P.A?Mono- window?Algorithm?for?RetrievingLand?Surface?Temperature?from?Landsat?TM?Data?and?Its?Application

When moisture content at 0.4～2.0g/cm ²The time:

τ ₃₁＝0.99513-0.0808w

τ ₃₂＝0.99376-0.11369w

When moisture content at 2～4.0g/cm ²The time:

τ ₃₁＝1.08692-0.12759w

τ ₃₂＝1.07900-0.15925w

When moisture content at 4.0～6.0g/cm ²The time:

τ ₃₁＝1.07268-0.12571w

τ ₃₂＝0.93821-0.12613w

The formula that calculates atmosphere vapour content is:

w＝[(α-lnT _w)/β] ²

Wherein w is the atmosphere vapour content of pairing each the pixel point of monitored farmland massif; T _wBe the ratio of the ground surface reflectance of the reflectivity of pairing each the pixel point of monitored farmland massif MODIS image the 18th wave band and pairing each the pixel point of the 2nd wave band, α, β are parameters, get α=0.02 respectively, β=0.651;

3) calculating of crop water supply index:

$\mathrm{VSWI}=\frac{\mathrm{EVI}}{T_s}$

VSWI is the crop water supply index of pairing each the pixel point of monitored farmland massif in the formula; EVI is the enhancing vegetation index of each pixel point of the pairing MODIS image of monitored farmland massif; T _sSurface temperature for pairing each the pixel point of remote sensing image of monitored farmland massif;

4) crop water supply index is carried out standardization:

SDI＝(VSWI-VSWI _d)/(VSWI _w-VSWI _d)×100％

SDI is the water supply index of each pixel point crop of the pairing remote sensing image of monitored farmland massif after the standardization in the formula, gets 0～100%, and wherein SDI=0 represents severe drought, and SDI=100% represents very moistening; VSWI _dAnd VSWI _wWhen being respectively the non-irrigateiest and the crop water supply index when the most moistening; The classification step-length of EVI can be made as d, when d=0.05, and when the temperature space of suitable plant growth is 20 ℃～45 ℃, VSWI _d=(n * d)/45, VSWI _w=(n * d)/20, n is for strengthening the number of vegetation index step-length, the positive integer of n 〉=1.For farmland ecosystem, VAWI _dAnd VSWI _wValue under the temperature of suitable growth is respectively:

5) calculating of precipitation anomaly index

$\mathrm{SRI}=\frac{R}{{2R}_w}\×100\%$

SRI is the precipitation anomaly index of monitored each pixel point of farmland massif in the formula, and SRI is big more moistening more; R is for working as the ten days rainfall amount; R _wWhen ten days rainfall amount mean value, get nearest 10 years for for many years in the method when ten days rainfall amount mean value.As R＞2R _wThe time for extremely moistening, get SRI=100%, R＞R _wThe time be normal, get SRI=50%, R is less than an average for many years half, is arid extremely, this moment, the SRI value was 0～25%.

Arid is a rainless meteorologic phenomena of long period, and a ten days rainlessly arid might not occur with interior, and the rainless of above time in a ten days just can produce arid problem, therefore the rainfall amount that needs to consider simultaneously early stage.This method has been considered the precipitation affects in nearest 8 ten days, thereby calculates comprehensive precipitation anomaly index, and its formula is:

SMRI＝A ₀×SRI ₀+A ₁×SRI ₁+A ₂×SRI ₂+A ₃×SRI ₃+…+A ₈×SRI ₈

6) determine agriculture damage caused by a drought index

DI＝B ₁×SDI+B ₂×SMRI

DI is the agriculture damage caused by a drought index of pairing each the pixel point of remote sensing image of monitored farmland massif in the formula, and it is the coupling of crop water supply exponential sum rainfall anomaly index, gets 0～100%, and DI=0 represents very arid, and DI=100% represents very moistening; SDI is a standardization water supply index, B ₁Be its weight, get 0.6; SMRI is a drought index of considering rain factor of many ten days, B ₂Be its weight, get 0.4; B ₁, B ₂The weight value be that value according to SDI conforms to the most with actual conditions.According to the rank criteria for classifying, judge the degree of arid then;

Herein, 1%≤DI≤15% drought of attaching most importance to, 15%＜DI≤30% is middle drought, and 30%＜DI≤50% is light drought, and 50%＜DI≤70% be that normally 70%＜DI≤100% is moistening.

The invention has the advantages that:

Utilize the MODIS remotely-sensed data that draught monitor is carried out in the farmland in big zone.In monitoring, model parameter is improved, adopted enhancing vegetation index (EVI) to replace normalized differential vegetation index (NDVI) that crop water supply index is calculated.NDVI and EVI are the vegetation indexs that Moderate Imaging Spectroradiomete data (MODIS) are selected for use, and EVI improves as the succession of NDVI with to some defective of NDVI, overall treatment Soil Background, atmospheric noise and ruddiness saturation problem.And the calculating of NDVI is the non-linear stretching near infrared and red spectral band, consequently strengthened low value part, suppressed high value part, and it is saturated that signal is appearred in high vegetation-covered area easily.It is easily saturated that EVI has then overcome the high areal coverage of vegetation, and the low areal coverage of vegetation is subjected to soil vegetative cover to influence bigger shortcoming.The present invention utilizes remote sensing technology to carry out the agricultural drought dynamic monitoring, and through practical application check, this method is easy, efficient, easy operating, result are accurate, can be widely used among the agricultural drought monitoring of China;

Embodiment

With the monitoring of North China's agricultural drought is example.Study area mainly comprises Hebei, Beijing, Tianjin, Shandong, Henan, Jiangsu and Anhui Province.This district is the main farming region of China, and major weather disaster is an arid, its occurrence frequency height, and longer duration seriously influences national economy.What local area often occurred is spring drought; Therefore the damage caused by a drought in North China April has been carried out study on monitoring.

At first select corresponding MODISL1B data according to the study area scope, (in remotely-sensed data softwares such as ERDAS or ENVI, handle) according to a conventional method the remotely-sensed data of selected wave band is carried out unified pre-service, utilize the remote sensing software data to carry out geometry correction, cloud detection and projection conversion, so that further mate calculating.Utilize each corresponding wave band data to calculate EVI, the T of each pixel point of monitored area respectively then _sWith crop water supply index.And select its pairing VSWI according to the size of the EVI value of monitored district each pixel point of remote sensing image _dAnd VSWI _wValue is carried out standardization to crop water supply index.Utilize the comprehensive rainfall anomaly of the rainfall data computation precipitation anomaly exponential sum index in the upper, middle and lower ten days in April of each meteorological site of study area; In ArcGIS software,, obtain and to delete trrellis diagram with the precipitation anomaly index of Remote Sensing Images Matching comprehensive precipitation anomaly exponential interpolation.At last according to the standardized crop water supply of precipitation anomaly exponential sum Index for Calculation agricultural damage caused by a drought index.

A kind of agricultural arid monitoring method based on the MODIS data mainly is to inquire into a kind of objective, dynamic, real-time, accurate, monitoring method of being easy to realize by the remote sensing technology means on big regional scale.Adopt in the method based on the agricultural arid monitoring model of vegetation index and surface temperature and monitor, in the process of monitoring the model Ecological Parameter is improved, the relevant parameter of the face of land damage caused by a drought of explaining is determined as follows:

Spectral reflectance data based on each wave band pixel of the pairing MODIS of monitored farmland massif;

1) calculating of vegetation index:

$\mathrm{EVI}=G\×\frac{{\mathrm{\ρ}}_{\mathrm{NIR}}-{\mathrm{\ρ}}_{\mathrm{Red}}}{{\mathrm{\ρ}}_{\mathrm{NIR}}+C_1\×{\mathrm{\ρ}}_{\mathrm{Red}}-C_2\×{\mathrm{\ρ}}_{\mathrm{Blue}}+L}$

EVI is for strengthening vegetation index, ρ _NIR, ρ _RedAnd ρ _BlueBe respectively the spectral reflectivity of near-infrared band, red spectral band and the blue wave band pixel of the pairing MODIS of monitored farmland massif; L is that background is adjusted item; C ₁And C ₂Be fitting coefficient; G is a gain factor; When calculating MODIS-EVI, L=1, C ₁=6, C ₂=7.5, G=2.5; Calculate the EVI data of pairing each the pixel point of remote sensing image of monitored farmland massif respectively;

2) Surface Temperature Retrieval

Surface temperature is a basic parameter of agricultural drought disaster monitoring, its computing method adopt (Qin Zhihao such as Qin Zhi person of outstanding talent, Gao Maofang, Qin Xiaomin etc., surface temperature remote sensing inversion method in the agricultural drought disaster monitoring-with MODIS data instance [J]. the disaster journal, 2005,14 (4): the 64-71.) window algorithm of splitting based on middle large scale MODIS data of Ti Chuing, its computing formula is:

T _s＝A ₀+A ₁T ₃₁-A ₂T ₃₂

A ₀＝E ₁a ₃₁-E ₂a ₃₂

A ₁＝1+A+E ₁b ₃₁

A ₂＝A+E ₂b ₃₂

A＝D ₃₁/E ₀

E ₁＝D ₃₂(1-C ₃₁-D ₃₁)/E ₀

E ₂＝D ₃₁(1-C ₃₂-D ₃₂)/E ₀

E ₀＝D ₃₂C ₃₁-D ₃₁C ₃₂

C _i＝ε _iτ _i

D _t＝[1+(1-ε _i)τ _i]

T in the formula _sBe surface temperature (K), T ₃₁And T ₃₂Be respectively the brightness temperature of MODIS the 31st and 32 wave bands, A ₀, A ₁And A ₂Be the parameter of splitting window algorithm, a ₃₁, b ₃₁, a ₃₁And b ₃₂Be constant, the desirable a of difference in surface temperature 0-50 ℃ scope ₃₁=-64.60363, b ₃₁=0.440817, a ₃₂=-68.72575, b ₃₂=0.473453; The calculating of other intermediate parameters sees the division window algorithm (Qin Zhihao at the MODIS data such as Qin Zhihao for details, Gao Maofang, Qin Xiaomin etc., the surface temperature remote sensing inversion method in the agricultural drought disaster monitoring-with MODIS data instance [J]. the disaster journal, 2005,14 (4): 64-71.); Utilize formula to calculate the T of pairing each the pixel point of remote sensing image of monitored farmland massif respectively then _sData;

Wherein i is meant the 31st and 32 wave bands of the pairing MODIS image of monitored farmland massif, is respectively i=31 or 32; τ _iBe the atmospheric transmittance of the wave band i of pairing each the pixel point of remote sensing image of monitored farmland massif, ε _iIt is the face of land emissivity of the wave band i of pairing each the pixel point of remote sensing image of monitored farmland massif.

ε _i＝ε _iw+P _vR _vε _iv+(1-P _v)R _sε _is

ε _Iw, ε _IvAnd ε _IsBe respectively water body, vegetation and the exposed soil of pairing each the pixel point of remote sensing image of monitored farmland massif face of land emissivity, get ε respectively at wave band i wave band _31w=0.99683, ε _32w=0.99254, ε _31v=0.98672, ε _32v=0.98990, ε _31s=0.96767, ε _31s=0.97790; R _vAnd R _sBe respectively the vegetation of the pairing remote sensing image of monitored farmland massif and the radiation ratio of exposed soil, (Qin Z such as Qin Zhihao, Karnieli A.Progress in the Remote Sensing of Land SurfaceTemperature and Ground Emissivity Using NOAA-AVHRR[J] .InternationalJournal of Remote Sensing, 1999,20 (12): 2367-2397.), the R of calculating _v=0.99240, R _s1.00744.P _vBe the vegetation coverage of pairing each the pixel point of monitored farmland massif, estimate by vegetation index:

$p_v=\frac{\mathrm{NDVI}-{\mathrm{NDVI}}_s}{{\mathrm{NDVI}}_v-{\mathrm{NDVI}}_s}$

NDVI is the vegetation index of pairing each the pixel point of monitored farmland massif in the formula, NDVI _vAnd NDVI _sBe respectively the NDVI value of dense vegetation covering and complete exposed soil pixel, get NDVI usually _v=0.674, NDVI _s=0.039 (Sun Rui, Zhu Qijiang. the analysis [J] of the clean primary productivity of clean primary productivity model of vegetation and China. Beijing Normal University's journal (natural science edition), 1998,34:132-137.);

$\mathrm{NDVI}=\frac{B_2-B_1}{B_2+B_1}$

B in the formula ₁And B ₂It is respectively the reflectivity of MODIS image the 1st and 2 wave bands.

Present near linear relationship between atmospheric transmittance and the atmosphere vapour content, according to Qin Zhihao etc. (

Qin?Z，Karnieli?A，Berliner?P.A?Mono- window?Algorithm?for?RetrievingLand?Surface?Temperature?from?Landsat?TM?Data?and?Its?Application?to

When moisture content at 0.4～2.0g/cm ²The time:

τ ₃₁＝0.99513-0.0808w

τ ₃₂＝0.99376-0.11369w

When moisture content at 2～4.0g/cm ²The time:

τ ₃₁＝1.08692-0.12759w

τ ₃₂＝1.07900-0.15925w

When moisture content at 4.0～6.0g/cm ²The time:

τ ₃₁＝1.07268-0.12571w

τ ₃₂＝0.93821-0.12613w

The formula that calculates atmosphere vapour content is:

w＝[(α-lnT _w)/β] ²

Wherein w is the atmosphere vapour content of pairing each the pixel point of monitored farmland massif; T _wBe the ratio of the ground surface reflectance of the reflectivity of pairing each the pixel point of monitored farmland massif MODIS image the 18th wave band and pairing each the pixel point of the 2nd wave band, α, β are parameters, get α=0.02 respectively, β=0.651;

3) calculating of crop water supply index:

$\mathrm{VSWI}=\frac{\mathrm{EVI}}{T_s}$

VSWI is the crop water supply index of pairing each the pixel point of monitored farmland massif in the formula; EVI is the enhancing vegetation index of each pixel point of the pairing MODIS image of monitored farmland massif; T _sSurface temperature for pairing each the pixel point of remote sensing image of monitored farmland massif;

4) crop water supply index is carried out standardization:

SDI＝(VSWI-VSWI _d)/(VSWI _w-VSWI _d)×100％

SDI is the water supply index of each pixel point crop of the pairing remote sensing image of monitored farmland massif after the standardization in the formula, gets 0～100%, and wherein SDI=0 represents severe drought, and SDI=100% represents very moistening; VSWI _dAnd VSWI _wWhen being respectively the non-irrigateiest and the crop water supply index when the most moistening; The classification step-length of EVI can be made as d, when d=0.05, and when the temperature space of suitable plant growth is 20 ℃～45 ℃, VSWI _d=(n * d)/45, VSWI _w=(n * d)/20, n be for strengthening the number of vegetation index step-length, the positive integer of n 〉=1 (in this example n determine that by every square kilometre of monitoring farmland massif area be 1 pixel, n is the pixel number).For farmland ecosystem, VAWI _dAnd VSWI _wValue under the temperature of suitable growth is respectively:

5) calculating of precipitation anomaly index

$\mathrm{SRI}=\frac{R}{{2R}_w}\×100\%$

SRI is the precipitation anomaly index of monitored each pixel point of farmland massif in the formula, and SRI is big more moistening more; R is for working as the ten days rainfall amount; R ^wWhen ten days rainfall amount mean value, get nearest 10 years for for many years in the method when ten days rainfall amount mean value.As R＞2R ^wThe time for extremely moistening, get SRI=100%, R＞R ^wThe time be normal, get SRI=50%, R is less than an average for many years half, is arid extremely, this moment, the SRI value was 0～25%.

Arid is a rainless meteorologic phenomena of long period, and a ten days rainlessly arid might not occur with interior, and the rainless of above time in a ten days just can produce arid problem, therefore the rainfall amount that needs to consider simultaneously early stage.This method has been considered the precipitation affects in nearest 8 ten days, thereby calculates comprehensive precipitation anomaly index, and its formula is:

SMRI＝A ₀×SRI ₀+A ₁×SRI ₁+A ₂×SRI ₂+A ₃×SRI ₃+…+A ₈×SRI ₈

6) determine agriculture damage caused by a drought index

DI＝B ₁×SDI+B ₂×SMRI

DI is the agriculture damage caused by a drought index of pairing each the pixel point of remote sensing image of monitored farmland massif in the formula, and it is the coupling of crop water supply exponential sum rainfall anomaly index, gets 0～100%, and DI=0 represents very arid, and DI=100% represents very moistening; SDI is a standardization water supply index, B ₁Be its weight, get 0.6; SMRI is a drought index of considering rain factor of many ten days, B ₂Be its weight, get 0.4; B ₁, B ₂The weight value be that value according to SDI conforms to the most with actual conditions.According to the rank criteria for classifying, judge the degree of arid then;

Herein, 1%≤DI≤15% drought of attaching most importance to, 15%＜DI≤30% is middle drought, and 30%＜DI≤50% is light drought, and 50%＜DI≤70% be that normally 70%＜DI≤100% is moistening.

In actual applications,, need add up area suffered from drought by calculating each province, the different damage caused by a drought grades in district and the ratio of the suffering from drought evaluation degree (table 1) of suffering from drought to the damage caused by a drought result in order to provide the damage caused by a drought result to relevant department more intuitively.

Table 1: each province in April, North China agricultural damage caused by a drought monitoring result

From this month precipitation event, western part, this month Hebei and big the precipitation in Henan are more on the low side, because cold and heat air is all very active, so strong convective weather is very many, the middle ten days and the last ten days, continuous thunderstorm weather has appearred in some places in Jiangsu, Anhui.But find out roughly that by the website observation data the general not enough 20mm of most of precipitation region adds that antecedent precipitation is on the low side, strong wind weather is many, and temperature significantly gos up, and causes the soil loss of moisture very fast, the high temperature and little rainfall that continue make the arid sustainable development in these places, to the end of month drought-hit area enlarge to some extent.From the rainfall data, this monitoring method can be good at reflecting the basic trend of agriculture damage caused by a drought.

Claims (2)

1. agricultural arid monitoring method based on the MODIS data is based on the spectral reflectance data of each wave band pixel of the pairing MODIS of monitored farmland massif; It is characterized in that:

1) calculating of vegetation index:

$\mathrm{EVI}=G\×\frac{{\mathrm{\ρ}}_{\mathrm{NIR}}-{\mathrm{\ρ}}_{\mathrm{Red}}}{{\mathrm{\ρ}}_{\mathrm{NIR}}+C_1\×{\mathrm{\ρ}}_{\mathrm{Red}}-C_2\×{\mathrm{\ρ}}_{\mathrm{Blue}}+L}$

EVI is for strengthening vegetation index, ρ _NIR, ρ _RedAnd ρ _BlueBe respectively the spectral reflectivity of near-infrared band, red spectral band and the blue wave band pixel of the pairing MODIS of monitored farmland massif; L is that background is adjusted item; C ₁And C ₂Be fitting coefficient; G is a gain factor; When calculating MODIS-EVI, L=1, C ₁=6, C ₂=7.5, G=2.5; Calculate the EVI data of pairing each the pixel point of remote sensing image of monitored farmland massif respectively;

2) Surface Temperature Retrieval, its computing formula is:

T _s＝A ₀+A ₁T ₃₁-A ₂T ₃₂

A ₀＝E ₁a ₃₁-E ₂a ₃₂

A ₁＝1+A+E ₁b ₃₁

A ₂＝A+E ₂b ₃₂

A＝D ₃₁/E ₀

E ₁＝D ₃₂(1-C ₃₁-D ₃₁)/E ₀

E ₂＝D ₃₁(1-C ₃₂-D ₃₂)/E ₀

E ₀＝D ₃₂C ₃₁-D ₃₁C ₃₂

C _i＝ε _iτ _i

D _i＝[1+(1-ε _i)τ _i]

T in the formula _sBe surface temperature (K), T ₃₁And T ₃₂Be respectively the brightness temperature of MODIS the 31st and 32 wave bands, A ₀, A ₁And A ₂Be the parameter of splitting window algorithm, a ₃₁, b ₃₁, a ₃₁And b ₃₂Be constant, the desirable a of difference in surface temperature 0-50 ℃ scope ₃₁=-64.60363, b ₃₁=0.440817, a ₃₂=-68.72575, b ₃₂=0.473453; Utilize formula to calculate the T of pairing each the pixel point of remote sensing image of monitored farmland massif respectively then _sData;

Wherein i is meant the 31st and 32 wave bands of the pairing MODIS image of monitored farmland massif, is respectively i=31 or 32; τ _iBe the atmospheric transmittance of the wave band i of pairing each the pixel point of remote sensing image of monitored farmland massif, ε _iIt is the face of land emissivity of the wave band i of pairing each the pixel point of remote sensing image of monitored farmland massif.

ε _i＝ε _iw+P _vR _vε _iv+(1-P _v)R _sε _is

ε _Iw, ε _IvAnd ε _IsThe water body, vegetation and the exposed soil that are pairing each the pixel point of remote sensing image of monitored farmland massif are respectively got ε respectively at the face of land of i wave band emissivity _31w=0.99683, ε _32w=0.99254, ε _31v=0.98672, ε _32v=0.98990, ε _31s=0.96767, ε _31s=0.97790; R _vAnd R _sBe respectively the vegetation of the pairing remote sensing image of monitored farmland massif and the radiation ratio of exposed soil, the R of calculating _v=0.99240, R _s=1.00744.P _vBe the vegetation coverage of pairing each the pixel point of monitored farmland massif, estimate by vegetation index:

$p_v=\frac{{\mathrm{NDVI}-\mathrm{NDVI}}_s}{{\mathrm{NDVI}}_v-{\mathrm{NDVI}}_s}$

NDVI is the vegetation index of pairing each the pixel point of monitored farmland massif in the formula, NDVI _vAnd NDVI _sBe respectively the NDVI value of dense vegetation covering and complete exposed soil pixel, get NDVI usually _v=0.674, NDVI _s=0.039;

$\mathrm{NDVI}=\frac{B_2-B_1}{B_2+B_1}$

B1 and B2 are respectively the reflectivity of MODIS image the 1st and 2 wave bands in the formula.

Present near linear relationship between atmospheric transmittance and the atmosphere vapour content, relational expression is

When moisture content at 0.4～2.0g/cm ²The time:

τ ₃₁＝0.99513-0.0808w

τ ₃₂＝0.99376-0.11369w

When moisture content at 2～4.0g/cm ²The time:

τ ₃₁＝1.08692-0.12759w

τ ₃₂＝1.07900-0.15925w

When moisture content at 4.0～6.0g/cm ²The time:

τ ₃₁＝1.07268-0.12571w

τ ₃₂＝0.93821-0.12613w

The formula that calculates atmosphere vapour content is:

w＝[(α-lnT _w)/β] ²

Wherein w is the atmosphere vapour content of pairing each the pixel point of monitored farmland massif; T _wBe the ratio of the ground surface reflectance of the reflectivity of pairing each the pixel point of monitored farmland massif MODIS image the 18th wave band and pairing each the pixel point of the 2nd wave band, α, β are parameters, get α=0.02 respectively, β=0.651;

3) calculating of crop water supply index:

$\mathrm{VSWI}=\frac{\mathrm{EVI}}{T_s}$

VSWI is the crop water supply index of pairing each the pixel point of monitored farmland massif in the formula; EVI is the enhancing vegetation index of each pixel point of the pairing MODIS image of monitored farmland massif; T _sSurface temperature for pairing each the pixel point of remote sensing image of monitored farmland massif;

4) crop water supply index is carried out standardization:

SDI＝(VSWI-VSWI _d)/(VSWI _w-VSWI _d)×100％

SDI is the water supply index of each pixel point crop of the pairing remote sensing image of monitored farmland massif after the standardization in the formula, gets 0～100%, and wherein SDI=0 represents severe drought, and SDI=100% represents very moistening; VSWI _dAnd VSWI _wWhen being respectively the non-irrigateiest and the crop water supply index when the most moistening; The classification step-length of EVI can be made as d, when d=0.05, and when the temperature space of suitable plant growth is 20 ℃～45 ℃, VSWI _d=(n * d)/45, VSWI _w=(n * d)/20, n is for strengthening the number of vegetation index step-length, the positive integer of n 〉=1.For farmland ecosystem, VAWI _dAnd VSWI _wValue under the temperature of suitable growth is respectively:

5) calculating of precipitation anomaly index

$\mathrm{SRI}=\frac{R}{{2R}_w}\×100\%$

SRI is the precipitation anomaly index of monitored each pixel point of farmland massif in the formula, and SRI is big more moistening more; R is for working as the ten days rainfall amount; R _wWhen ten days rainfall amount mean value, get nearest 10 years for for many years in the method when ten days rainfall amount mean value.As R＞2R _wThe time for extremely moistening, get SRI=100%, R＞R _wThe time be normal, get SRI=50%, R is less than an average for many years half, is arid extremely, this moment, the SRI value was 0～25%.

Arid is a rainless meteorologic phenomena of long period, and a ten days rainlessly arid might not occur with interior, and the rainless of above time in a ten days just can produce arid problem, therefore the rainfall amount that needs to consider simultaneously early stage.This method has been considered the precipitation affects in nearest 8 ten days, thereby calculates comprehensive precipitation anomaly index, and its formula is:

SMRI＝A ₀×SRI ₀+A ₁×SRI ₁+A ₂×SRI ₂+A ₃×SRI ₃+…+A ₈×SRI ₈

6) determine agriculture damage caused by a drought index

DI＝B ₁×SDI+B ₂×SMRI

DI is the agriculture damage caused by a drought index of pairing each the pixel point of remote sensing image of monitored farmland massif in the formula, and it is the coupling of crop water supply exponential sum rainfall anomaly index, gets 0～100%, and DI=0 represents very arid, and DI=100% represents very moistening; SDI is a standardization water supply index, B ₁Be its weight, get 0.6; SMRI is a drought index of considering rain factor of many ten days, B ₂Be its weight, get 0.4; B ₁, B ₂The weight value be that value according to SDI conforms to the most with actual conditions.According to the rank criteria for classifying, judge the degree of arid then;

Herein, 1%≤DI≤15% drought of attaching most importance to, 15%＜DI≤30% is middle drought, and 30%＜DI≤50% is light drought, and 50%＜DI≤70% be that normally 70%＜DI≤100% is moistening.

2. according to the described method of claim 1, it is characterized in that: obtain EVI, T _s, VSWI and SDI parameter, the remotely-sensed data of selected wave band is carried out unified pre-service, mainly comprise and carry out radiant quantity calculating, geometry correction, cloud detection, projection conversion according to a conventional method, so that further mate calculating; When calculating rainfall anomaly index, at first the rainfall data should be carried out interpolation processing, obtain precipitation anomaly figure; The unified raster data that changes into after interpolation is finished, so as with the remotely-sensed data comparative analysis.