CN108733619A - Global arbitrary forest bottom class growth prediction model quantitative estimation method - Google Patents
Global arbitrary forest bottom class growth prediction model quantitative estimation method Download PDFInfo
- Publication number
- CN108733619A CN108733619A CN201810472607.8A CN201810472607A CN108733619A CN 108733619 A CN108733619 A CN 108733619A CN 201810472607 A CN201810472607 A CN 201810472607A CN 108733619 A CN108733619 A CN 108733619A
- Authority
- CN
- China
- Prior art keywords
- condition
- bottom class
- forest
- model
- class
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Algebra (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Global arbitrary forest bottom class growth prediction model quantitative estimation method.A kind of method of multivariate regression models quantitative estimation bottom class growth, it is characterised in that：It chooses and the factor (Type of Forest Land being affected is estimated to bottom class's growing state, geographical location, Astronomical Condition, weather conditions, edaphic condition, solar term condition) multivariate regression models is established, inverting resolving is carried out to model using bottom class's data of different condition factor, obtains bottom class's growing state prediction based on different woodland type and the regression relation between each impact factor.Later, specified bottom class is quickly estimated using corresponding model, growth prediction situation is calculated.
Description
One, technical field
The present invention relates to the quantitative estimation methods of global arbitrary forest bottom class growth prediction model, in particular with big data
The pattern of prediction establishes forest bottom class growing state and geographical location, Astronomical Condition, weather conditions, edaphic condition, solar term condition
Etc. a series of conditions relationship, carry out multiple regression analysis in conjunction with the remote sensing informations data such as RS, UAV, and then predict that the whole world is arbitrary
The method of forest bottom class growing state.
Two, technical background
Forest is the biocoene based on xylophyta, be arbor and the other plants concentrated, animal, microorganism and
Interdependence mutually restricts between soil, and and environmental interaction, to the totality of the ecosystem formed.Forest makees
For " lung of the earth ", there is abundant species, complicated structure, diversified function；Forest is maximum on the earth simultaneously
Terrestrial ecosystems, and forest reserves is not only one country forest reserves total scale of reflection and horizontal basic finger
One of mark, and the abundant degree of the reflection forest reserves, the important evidence for weighing forest ecological environment quality.
The forest reserves are one of most important resources on the earth, are the key that forestry sustainable developments, how to measure forest
Current resources situation monitors the variation of its dynamical system, and quantity, quality, distribution and health of the assay forest reserves etc. are that we work as
Modern hot issue of interest.With the development of deep space satellite, remote sensing technology has become the main support of forest resource monitoring
Means, the coverage information of remote sensing image with having recorded earth's surface truly class, different types of ground objects due to its spectral characteristic difference, because
This brightness value on image is also different, this allows to distinguish different types of ground objects according to remote sensing image.
In the 1970s, remote sensing technology is introduced into China, with the tremendous development of remote sensing cause, remote sensing technology quickly by
Applied to field of forestry, the growth prediction of forest bottom class is also at the hot issue of research.Academician Qian Xuesen is once in 20th century 60
As long as age proposition possesses sufficient data, then for agricultural plant growth the case where can necessarily carry out Accurate Prediction.With
The arriving in big data epoch, the abundant development of remote sensing technology, this is imagined becomes possibility now.
Technical way：Using Forest Types and site quality combination Landsat satellites, SPOT satellites, IKONOS is defended
The satellite remote-sensing image data progress forest of star, China-Brazil Earth Resources Satellite, a series of home and abroads such as No. three satellites of resource is small
Class's growth prediction scale-model investigation, using the round-the-clock of satellite, multidate, at many levels, and the features such as data acquisition amount is big, remote sensing image
Data combination sample-plot survey data, combining geographic location, Astronomical Condition, weather conditions, edaphic condition, solar term condition etc. use
It is small that the multiple regressions method of estimation such as one-variable linear regression and successive Regression, offset minimum binary, ridge estimaion establishes the arbitrary forest in the whole world
Class's growth prediction model.
Three, invention content
For the limitation for overcoming traditional forest bottom class growth model to calculate, realize that quickly and efficiently the evaluation whole world is arbitrary gloomy
The growing state of woods bottom class, the object of the present invention is to provide a kind of methods of multivariate regression models quantitative estimation bottom class growth.
BROAD SUMMARY：
It chooses and the factor (Type of Forest Land, geographical location, Astronomical Condition, the weather that are affected is estimated to bottom class's growing state
Condition, edaphic condition, solar term condition) multivariate regression models is established, model is carried out using bottom class's data of different condition factor
Inverting resolves, and obtains bottom class's growing state prediction based on different woodland type and the recurrence between each impact factor is closed
System.Later, specified bottom class is quickly estimated using corresponding model, growth prediction situation is calculated.
This invention has the following advantages compared with the conventional method：
(1) measuring and calculating of traditional bottom class growth model is compared, this invention is existed using big data, and integral data handles data
The advantage of aspect, it is round-the-clock in conjunction with satellite remote sensing, the characteristics of multidate, for example, in December, 2016 transmitting high scape No.1, can be with
Reach 0.5 meter of resolution ratio, and single satellite can acquire 700,000 square kilometres daily.In the whole world Anywhere, it can be achieved that every
Its observation is primary.Corresponding classification prediction model is applied mechanically using these advantages to estimate specified forest bottom class growing state,
It is more convenient, reduce and manually calculates workload on the spot.
(2) all data whole presence server ends, can obtain, and quickly calculate in real time, as long as anyone input must
The information condition wanted can obtain the forest bottom class growth prediction model of corresponding area.
(3) the considerations of proposing Type of Forest Land and site quality, and consider geographical location, Astronomical Condition, weather conditions, soil
Condition, the various condition elements such as solar term condition, the classification regression model established in conjunction with remotely-sensed data are compared unified regression model and are had
There is a higher estimation precision, and a kind of new thinking and resolving approach are provided for bottom class's growth prediction measuring and calculating.
Four, it illustrates：
The arbitrary forest bottom class growth prediction model foundation flow chart in the whole world Fig. 1
Five, specific implementation mode：
The present invention provides a kind of new resolving thinkings and pattern, specifically：
(1) means such as remote sensing satellite (RS), unmanned plane (UAV), assize super-station instrument, 3D electronic angle gauges, micro- sample plot method are utilized
Land occupation condition similar with equipment realization and crop type observation, observation element have：1, forestry resource survey data；2, bottom class's classification, area, it is raw
Long situation.On the basis of sufficient data acquire, forest bottom class growing state sample database is established.
(2) choose influences relatively large factor in forestry resource survey data on forest bottom class growing state establishes multiple regression
Analysis model：
In the model：Y is the subcompartment's stock volume (m of unit area3/hm2)；
S is subcompartment area (hm2)；
a0For correction value；
The model is divided into geographic element G, and sun element T, weather conditions C edaphic conditions S and five aspects of solar term condition are right
It is predicted in forest bottom class growing state
1. geographic element G, according to the geographical location residing for forest bottom class, we can be expressed with following 5 elements：H,X,
Y、αi、βi, they are respectively in geographic element：Height above sea level (elevation), plan-position X, Y, the gradient (0 °~90 °), slope aspect；
2. sun element T：Plant growth indispensable element when the irradiation of sunlight, the major influence factors of the sun
For：Sun altitude γ, when to τ；
3. climatic elements C：Weather conditions also restrict the growth of plant, and major influence factors can be summarized as temperature t,
HumidityRainfall k；
4. edaphic condition S：The condition of soil, which affects, can be summarized as three most important factor λ, h, σ, respectively native
Earth type, thickness of soil, soil nutrient parameter；
5. solar term condition：Solar term are a kind of supplement calendars for guiding agricultural production that ancient Chinese is concluded, it can be preferably
The sunny operation of reflection period.According to 24 solar term heats computation year, corresponding affecting parameters ε is taken according to model
A1, a2 in formula ..., the coefficient that a13 is each impact factor；
(3) it brings formula 1 into respectively using different bottom class's sample datas and carries out multiple regression analysis, it is polynary according to Fig. 1 flows
The step of regression analysis, is as follows：
Regression equation is investigated, dependent variable is forest bottom class growing state value Y, and dependent variable is geographic element G (H, X, Y, αi、
βi), sun element T (γ, τ), climatic elementsEdaphic condition S (λ, h, σ), solar term condition (ε), totally 13, accidentally
Difference is a0 dependent variable coefficients a1, a1 ..., a13
Multivariable regressive analysis model can be established：
If stochastic variable Y (forest bottom class growing state value) and common variables (including geographic element G (H, X, Y, αi、βi),
Sun element T (γ, τ), climatic elementsEdaphic condition S (λ, h, σ), solar term condition (ε)) between exist it is linear
Relationship, and formula can be expressed as：
This formula is multiple linear regression equations of the stochastic variable Y about common variables, and a0 is regression constant, a1, a2 ...
A13 is partial regression coefficient, in order to estimate to obtain unknown parameter a0, a2 ... a13, enables geographic element G (H, X, Y, αi、βi), sun element
T (γ, τ), climatic elementsEdaphic condition S (λ, h, σ) is respectively x1, x2 ... x13, former multiple regression equation
It can be write as：
Y=β0+β1x1+…+β13x13 (2.2)
Wherein：
And polynary empirical regression equation can be obtained：
To (x1,x2…x13, y) and n times independent observation is carried out, obtain the sample set that sample size is n：
(xi1,xi2…xi13, y) and (i=1,2 ... ..., n)
For formula 2.1, estimated value b0,b1……b13Should make observation yiWith corresponding regressand valueSum of squares of deviations is most
Small β0,β1……β13Value, i.e.,
It is minimum
A system of linear equations is obtained after seeking extreme value：
l11β1+l12β2+…+l1kβk=l1y
l21β1+l22β2+…+l2kβk=l2y
……
lk1β1+lk2β2+…+lkkβk=lky
Wherein：K=13
It is β that it, which is solved,0,β1……β13Least-squares estimation, be to acquire linear regression model (LRM)：
For multiple regression equation hypothesis testing：
Assuming that H0:b0,b1……b13It is 0, then having：
Thus F is calculated0, and significance probability Sig.=P (F are calculated by its corresponding F distribution>F0), if Sig.<0.1,
Regression equation significant effect is then thought, otherwise it is assumed that it is not notable.
The significance test of single coefficient：
Assuming that H0:bi=0, then having：
Wherein ciiFor the inverse matrix diagonal element of normal equation group coefficient matrix, T is calculatedi, and by its corresponding T distributions meter
Calculate significance probability Sig.=P (t=Ti), if Sig.<0.05, then it is assumed that coefficient biIt is not significantly 0, it is thus regarded that xiHave to y aobvious
Writing influences, otherwise it is assumed that coefficient biIt is approximately 0, it is thus regarded that xiY is had no significant effect.
(4) multiple linear regression analysis and inspection can be carried out according to above principle：
1. first, whether there is line between each influence factor (explanatory variable) and Y value (dependent variable) using sample set
Sexual intercourse carries out significance test to each influence factor, and the hypothesis for being 0 for coefficient carries out hypothetical inspection, is less than 0.1
Refuse null hypothesis, you can to think that the statistic is notable.For error coefficient a0F inspections are carried out, it is equally small in significance
When 0.1, refuse null hypothesis, it is believed that error coefficient a0Significantly.
2. due to there is interaction between each influence factor, phase of each independent variable respectively between dependent variable is only seen
Relationship number can not reflect the truth between each variable completely.So choosing PARCOR coefficients, control is other
The influence of correlativity between two variables of variable pair.For example, the height above sea level in selection geographic element, height above sea level H is for forest for analysis
The influence of bottom class's growing state, is examined by T, and the significance of T values is approximately 0, it is believed that height above sea level H is for forest bottom class
The influence of growing state is notable；PARCOR coefficients are 0.162, it is believed that variable and dependent variable are relevant.It similarly, can be with
Tentatively judging its dependent variable also influences significantly, can enter model.
3. utilizing spss, the statistical analysis softwares such as R language calculate multivariate regression models ginseng in conjunction with the data in sample set
Number, can obtain a0、a1……、a13Significance test is carried out etc. the value of each parameter, and to these parameters, is examined by T, if
The significance of T values is approximately 0, it is believed that corresponding parameter meets influence factor for forest bottom class growing state
Influence degree.The related coefficient for calculating multivariate regression models, verifies the reasonability of parameters.
4. calculating the residual error of multivariate regression models according to obtained parameter, the standardized residual figure and residual error of residual error are investigated
The normality and equal variance of item show that standardized residual is basic between match value if data are uniformly distributed in 0 horizontal line both sides
There is no trend sexual intercourse, illustrate that the exponential model established has reasonability.
5. by analyzing above, the reasonability of model can be verified, next according to obtained multivariate regression models into
Row prediction and control.
(5) it for the forest bottom class in the global range of certain selected as research object, collects its corresponding geography and wants
Element (H, X, Y, αi、βi), sun element T (γ, τ), climatic elementsEdaphic condition S (λ, h, σ), solar term condition
(ε) can solve the Y value in equation using above-mentioned multiple regression equation, can be for the growth feelings of the bottom class according to Y value
Condition effectively predict and assess.It is also possible to according to prediction as a result, timely progress disease protection, scientifically and rationally
Guidance is arranged production, and ensures the growth of forest.
Place is not described in detail for this specification, is the known technology of those skilled in the art of the present technique.
Claims (1)
1. global arbitrary stigma Forest Growth prediction model, it is characterised in that：It chooses and is affected to bottom class's growing state estimation
Factor (Type of Forest Land, geographical location, Astronomical Condition, weather conditions, edaphic condition, solar term condition) establish multiple regression mould
Type carries out inverting resolving to model using bottom class's data of different condition factor, obtains small based on different woodland type
Regression relation between class's growing state prediction and each impact factor
One, choose influences relatively large factor in forestry resource survey data on forest bottom class growing state establishes multiple regression analysis
Model：
In the model：Y is the subcompartment's stock volume (m of unit area3/hm2)；
S is subcompartment area (hm2)；
a0For correction value；
The model is divided into geographic element G, and sun element T, weather conditions C edaphic conditions S and five aspects of solar term condition are for gloomy
Woods bottom class growing state is predicted
1. geographic element G, according to the geographical location residing for forest bottom class, we can be expressed with following 5 elements：H,X,Y,αi、
βi, they are respectively in geographic element：Height above sea level (elevation), plan-position X, Y, the gradient (0 °~90 °), slope aspect；
2. sun element T：Plant growth indispensable element when the irradiation of sunlight, the major influence factors of the sun are：Too
Positive elevation angle γ, when to τ；
3. climatic elements C：Weather conditions also restrict the growth of plant, and major influence factors can be summarized as temperature t, humidityRainfall k；
4. edaphic condition S：The condition of soil, which affects, can be summarized as three most important factor λ, h, σ, respectively soil class
Type, thickness of soil, soil nutrient parameter；
5. solar term condition：Solar term are a kind of supplement calendars for guiding agricultural production that ancient Chinese is concluded, can be preferably anti-
The period for mirroring sun operation takes corresponding affecting parameters ε according to 24 solar term heats computation year according to model
A1, a2 in formula ..., the coefficient that a13 is each impact factor；
Multiple regression analysis is carried out using formula 1 is brought into respectively using different bottom class's sample datas, calculate a0, a1, a1 ...,
A13 coefficients
Two, for the forest bottom class in the global range of certain selected as research object, collect its corresponding geographic element (H,
X、Y、αi、βi), sun element T (γ, τ), climatic elementsEdaphic condition S (λ, h, σ), solar term condition (ε),
Using multiple regression equation described in formula 1, the Y value in equation can be solved, it, can be for the growth of the bottom class according to Y value
Situation effectively predict and assess.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810472607.8A CN108733619A (en) | 2018-05-17 | 2018-05-17 | Global arbitrary forest bottom class growth prediction model quantitative estimation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810472607.8A CN108733619A (en) | 2018-05-17 | 2018-05-17 | Global arbitrary forest bottom class growth prediction model quantitative estimation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108733619A true CN108733619A (en) | 2018-11-02 |
Family
ID=63938456
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810472607.8A Pending CN108733619A (en) | 2018-05-17 | 2018-05-17 | Global arbitrary forest bottom class growth prediction model quantitative estimation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108733619A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135913A (en) * | 2019-05-20 | 2019-08-16 | 智慧足迹数据科技有限公司 | Training method, shop site selecting method and the device of shop site selection model |
| CN110853699A (en) * | 2019-10-30 | 2020-02-28 | 北京林业大学 | Method for establishing single-tree growth model under large-area condition |
| CN111415065A (en) * | 2020-02-26 | 2020-07-14 | 广州地理研究所 | Mountain disaster ecological damage risk evaluation method based on action process |
| CN111563824A (en) * | 2020-06-05 | 2020-08-21 | 西北农林科技大学 | Artificial forest land research system and method based on Chinese pine |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE19544112C2 (en) * | 1995-11-27 | 2001-10-18 | Claas Kgaa Mbh | Process for generating digital terrain relief models |
| CN104020274A (en) * | 2014-06-05 | 2014-09-03 | 刘健 | Method for remote sensing quantitative estimation on woodland site quality |
| CN105893737A (en) * | 2016-03-24 | 2016-08-24 | 北京林业大学 | Construction method of geological model-based forest growing stock estimation model |
-
2018
- 2018-05-17 CN CN201810472607.8A patent/CN108733619A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE19544112C2 (en) * | 1995-11-27 | 2001-10-18 | Claas Kgaa Mbh | Process for generating digital terrain relief models |
| CN104020274A (en) * | 2014-06-05 | 2014-09-03 | 刘健 | Method for remote sensing quantitative estimation on woodland site quality |
| CN105893737A (en) * | 2016-03-24 | 2016-08-24 | 北京林业大学 | Construction method of geological model-based forest growing stock estimation model |
Non-Patent Citations (2)
| Title |
|---|
| 吴达胜: "基于多源数据和神经网络模型的森林资源蓄积量动态监测", 《中国博士学位论文全文数据库 农业科技辑》 * |
| 王佳 等: "结合影像光谱与地形因子的森林蓄积量估测模型", 《农业机械学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135913A (en) * | 2019-05-20 | 2019-08-16 | 智慧足迹数据科技有限公司 | Training method, shop site selecting method and the device of shop site selection model |
| CN110853699A (en) * | 2019-10-30 | 2020-02-28 | 北京林业大学 | Method for establishing single-tree growth model under large-area condition |
| CN111415065A (en) * | 2020-02-26 | 2020-07-14 | 广州地理研究所 | Mountain disaster ecological damage risk evaluation method based on action process |
| CN111415065B (en) * | 2020-02-26 | 2023-05-12 | 广州地理研究所 | Mountain disaster ecological damage risk evaluation method based on action process |
| CN111563824A (en) * | 2020-06-05 | 2020-08-21 | 西北农林科技大学 | Artificial forest land research system and method based on Chinese pine |
| CN111563824B (en) * | 2020-06-05 | 2024-02-23 | 西北农林科技大学 | Artificial forest land research system and method based on Chinese pine |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Anderegg et al. | Hydraulic diversity of forests regulates ecosystem resilience during drought | |
| Feng et al. | Evaluation of sunshine-based models for predicting diffuse solar radiation in China | |
| Li et al. | Vegetation control on water and energy balance within the Budyko framework | |
| Bonan et al. | Reconciling leaf physiological traits and canopy flux data: Use of the TRY and FLUXNET databases in the Community Land Model version 4 | |
| Antonarakis et al. | Using Lidar and Radar measurements to constrain predictions of forest ecosystem structure and function | |
| Xu et al. | Estimating turbulent fluxes through assimilation of geostationary operational environmental satellites data using ensemble Kalman filter | |
| Meng et al. | Investigating spatiotemporal changes of the land-surface processes in Xinjiang using high-resolution CLM3. 5 and CLDAS: Soil temperature | |
| CN108733619A (en) | Global arbitrary forest bottom class growth prediction model quantitative estimation method | |
| Lucas et al. | An observational analysis of Southern Hemisphere tropical expansion | |
| Chen et al. | A column canopy‐air turbulent diffusion method for different canopy structures | |
| Luo et al. | Estimation of total cloud cover from solar radiation observations at Lake Rotorua, New Zealand | |
| Wang et al. | Integrating remotely sensed leaf area index and leaf nitrogen accumulation with RiceGrow model based on particle swarm optimization algorithm for rice grain yield assessment | |
| Li et al. | Estimation of evapotranspiration over the terrestrial ecosystems in China | |
| Sawada et al. | Simultaneous estimation of both hydrological and ecological parameters in an ecohydrological model by assimilating microwave signal | |
| Belviso et al. | Comparison of global climatological maps of sea surface dimethyl sulfide | |
| Petropoulos et al. | A global Bayesian sensitivity analysis of the 1d SimSphere soil–vegetation–atmospheric transfer (SVAT) model using Gaussian model emulation | |
| Deng et al. | Comparison of soil moisture products from microwave remote sensing, land model, and reanalysis using global ground observations | |
| Jung et al. | Comparison of the Penman‐Monteith method and regional calibration of the Hargreaves equation for actual evapotranspiration using SWAT-simulated results in the Seolma-cheon basin, South Korea | |
| McCabe et al. | Calibration of a land surface model using multiple data sets | |
| Cammalleri et al. | State and parameter update in a coupled energy/hydrologic balance model using ensemble Kalman filtering | |
| Li et al. | Production of a combined land surface data set and its use to assess land‐atmosphere coupling in China | |
| Vilasa et al. | Global soil moisture bimodality in satellite observations and climate models | |
| Hu et al. | Progress, challenges, and future steps in data assimilation for convection‐permitting numerical weather prediction: Report on the virtual meeting held on 10 and 12 November 2021 | |
| Yan et al. | Spectroscopy outperforms leaf trait relationships for predicting photosynthetic capacity across different forest types | |
| Möller et al. | Adjustment of regional climate model output for modeling the climatic mass balance of all glaciers on Svalbard |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181102 |