CN111783360B

CN111783360B - High-resolution land utilization and forest landscape process coupling simulation system and method

Info

Publication number: CN111783360B
Application number: CN202010639670.3A
Authority: CN
Inventors: 梁宇; 黄超
Original assignee: Institute of Applied Ecology of CAS
Current assignee: Institute of Applied Ecology of CAS
Priority date: 2020-07-06
Filing date: 2020-07-06
Publication date: 2024-04-16
Anticipated expiration: 2040-07-06
Also published as: CN111783360A

Abstract

The invention provides a high-resolution land utilization and forest landscape process coupling simulation system and method, and relates to the technical field of geography and ecology. The system comprises a land utilization change module and a forest succession module; the land utilization change module calculates land utilization pattern change under different social economy and natural conditions in the future by utilizing a self-adaptive inertial competition mechanism on the basis of a neural network model; the forest succession module is a space visual landscape model for simulating forest succession on a large space-time scale, and the influence of future climate change and land utilization change on the carbon reserves of the forest ecosystem in China can be evaluated more accurately. Compared with the existing ecological system process model, the model has the advantages of fewer required parameters, more convenient acquisition, higher operation efficiency and capability of directly using forest resource investigation data for simulation precision of the parameterized model and the verification model.

Description

High-resolution land utilization and forest landscape process coupling simulation system and method

Technical Field

The invention relates to the technical field of geography and ecology, in particular to a high-resolution land utilization and forest landscape process coupling simulation system and method.

Background

Land utilization changes are one of the most important factors affecting carbon circulation in land ecosystems, which is CO to the atmosphere ₂ The contribution of the increase in concentration is inferior to the combustion of fossil fuels (Foley et al 2005, houghton et al 2012). Studies have shown that 1850-1998 had 136.+ -. 5X 10 ⁹ t carbon is discharged from land ecosystem to atmosphere through land utilization change, and occupies CO increased in the current atmosphere ₂ 33% of the total (Watson et al 2000). Forest is the largest carbon reservoir of the terrestrial ecosystem, whose carbon reserves account for about 45% of the carbon reserves of the global terrestrial ecosystem, and plays an important role in the global carbon cycle (Pan et al 2011, reich 2011). The change of land utilization not only directly reduces the carbon reserves on the ground of the forest ecosystem (Baccii et al 2012), but also indirectly influences the content and distribution of the organic carbon in the soil by influencing the factors forming and converting with the organic carbon in the soil (Don et al 2011), and can influence the accumulation of the organic carbon and nitrogen in the soil by changing the decomposition rate of the organic matters in the soil, and the tiny fluctuation of the method can slow down or accelerate the CO in the atmosphere to a great extent ₂ The concentration was increased, changing the global terrestrial ecosystem carbon balance (Le Qu re et al 2018). Thus, land use changes are ecological to forests The influence of the carbon reserves of the system is paid special attention, is an important basis for forest ecological management, and has important guiding significance for the establishment of climate change slowing policies by the government of China.

The current method for researching the influence of land utilization change on the carbon reserves of the land ecological system mainly comprises an empirical model, an IPCC inventory method, an ecological process model and the like. The empirical model is to build a statistically relevant model through empirical relationships to quantify the effect of land use changes on the carbon balance of the terrestrial ecosystem (by et al 2012). The "Bookkeeping" model established by Houghton et al is the most common and widespread empirical statistical model currently studying land use changes for land ecological carbon balance (Houghton et al 1983). The model calculates regional carbon balance changes based on land use change rates and changes in vegetation and soil carbon reserves per unit area that are closely related to land use changes. The response curves of vegetation and soil carbon reserves to several typical land use changes are given in the "Bookkeeping" model: forest cutting, harvesting and afforestation of forest products, abandoned tillage and barren work, and barren and expansion of grasslands. The model is widely used for estimating the influence of land utilization change on the area and global scale on the carbon circulation of a forest ecosystem through continuous correction and perfection of a plurality of domestic and foreign scholars (Ge Quansheng et al 2008). However, land utilization change is a complex process, and a bookkeeping model takes vegetation and soil carbon reserves as key points of the whole model, and carbon emission caused by processes of soil respiration, oxidation and decay of remained vegetation residues and the like in the land utilization change process is not considered. In addition, different researches have differences on land utilization change rate calculation methods and carbon density data acquisition modes, so that the estimation results of the Bookkeeping model have larger uncertainty (Tang Rui and Peng Kaili 2018).

Although existing ecosystem process models can better simulate carbon balance of a forest ecosystem, most of the models cannot directly simulate future land utilization changes, and direct and indirect influences of the land utilization changes on the forest ecosystem are not fully considered. Therefore, building a coupling model of land use changes and ecosystems is a future development direction (Lapola and Priess 2009, tian Hanqin et al 2010). There are few models currently available to couple land use changes with forest ecosystems processes, the most representative models being the IMAGE (Leemans and Eickhout 2004,Lapola et al.2010) and LPJmL models (yglin et al 2010). The IMAGE model couples an ecological process model, land utilization and a socioeconomic model, and predicts future land utilization space patterns by using socioeconomic data such as population growth, economy, energy supply, demand and the like as a drive, so as to simulate land utilization changes and carbon emission caused by the land utilization changes (Leemans et al 1996). The LPJmL model uses 0.5 ° grid units to simulate the effect of land demand and land utilization grid changes under different policy targets on the carbon circulation of the global forest ecosystem (Lapola et al 2009). The IMAGE and LPJmL models can use the change data of the global land to simulate the global and intercontinental scales, but because the simulation pixels of the 2 models are relatively large (0.5-2.5 degrees), the simulation of the carbon circulation process of the forest ecosystem is too simplified, so that the model estimation results are greatly different from the observation data, and the large difference exists between different model results (Chen Ansheng and Tian Hanqin 2007). In addition, the existing ecological process model cannot directly simulate the influence of land utilization changes such as forestation barren, forestation and urbanization on vegetation succession of a forest ecological system, and the influence of the land utilization changes on carbon reserves of the forest ecological system can be underestimated.

Disclosure of Invention

Aiming at the problem of influence research on carbon reserves of a forest ecological system by the existing land utilization change, the invention provides a high-resolution land utilization and forest landscape process coupling simulation system and method;

the technical scheme adopted by the invention is as follows:

on one hand, the invention provides a high-resolution land utilization and forest landscape process coupling simulation system, which comprises a land utilization change module and a forest succession module;

the land utilization change module calculates land utilization pattern change under different social economy and natural conditions in the future by utilizing a self-adaptive inertial competition mechanism on the basis of a neural network model; the forest succession module is a space visual landscape model for simulating forest landscape change on a large space-time scale, different stages of forest stand development are defined according to 4 GSO thresholds, and the module simulates the influence of forest succession, seed diffusion, forest fire, harvesting, wind fall, plant diseases and insect pests and combustible treatment processes on the structure and functions of a forest ecological system; the module respectively simulates the influence of tree species updating, growth, death and forest stand state change on a forest ecological system on tree species, forest stand and landscape scale; simulating the updating, growing, dying and propagating processes of single trees on the tree species scale; simulating species colonisation, self-sparsity, under-forest regeneration and maturation processes on a stand scale; on the landscape scale, simulating forest succession, seed propagation, competition among seeds, forest fire, harvesting, wind fall, plant diseases and insect pests and combustible treatment processes by tracking the existence and disappearance of tree species;

The 4 GSO thresholds include: GSO (GSO) ₁ Occupying part of the growth space stage, and all the reached tree species are germinated and colonised except the most negative-tolerant tree species; GSO (GSO) ₁ =1, canopy-closed stage, when only the most yin-tolerant tree species germinate and colonise; GSO (GSO) ₃ Tree species already occupy the whole growth space, and the vigorous competition leads to the inability of all tree species to colonise; GSO (GSO) _thinning From the sparsity stage, the stand density begins to decrease.

The growth of the tree is determined by a growth space, and the growth space refers to an area capable of providing nutrients, water and illumination for plant growth; in a limited growth space, different tree species compete for more growth space, thereby causing death of tree species and colonization of seedlings; in the forest succession module, the growth space is calculated by a forest Stand Density Index (SDI), a breast Diameter (DBH) and a stump density (TPA), wherein the forest stand density index is expressed as follows:

wherein: TPA is the standing tree density, i.e., the total number of all living standing trees with a height of more than 5cm per unit area; DBH is breast diameter, which refers to the diameter of the trunk of standing tree at 1.3m from the ground surface;

the propagation of the tree is that the tree is propagated in a root tiller mode after the tree death caused by forest fire or harvesting or the physiological damage caused by drought; in the forest succession module, the potential germination number and the seedling colonisation number are obtained by estimating the number of trees at each age level, and the potential germination number PPS of the seedlings is calculated as follows:

PPS＝NDT×VP

Wherein, NDT refers to the death plant number of a certain tree species, VP refers to the asexual reproduction probability of the tree species;

the seed propagation is a spatial process, and the seed propagation probability P is obtained by the following formula:

P＝e ^-b (x/MD)ED<x<MD

wherein MD is the maximum propagation distance of a certain tree species; x refers to the distance from the seed source; b is a seed propagation parameter; ED is the effective propagation distance of the seeds of the tree species; e refers to an exponential function based on a natural constant e;

the self-sparsity is a process that different tree individuals compete for light, water, heat and nutrition conditions, the difference among the individuals is larger and larger along with the growth, and the inferior tree individuals die, namely the individual density is gradually reduced along with the succession of forests; 3/2 self-thinning and occupied Growth Space (GSO) was used to simulate the sparseness process during forest distribution as follows:

in the formula, DBH _v And NT _V Mean chest diameter and plant number of a certain tree species at the V-th diameter level; 10inch refers to a standard tree with a chest diameter of 10 inches (25.4 cm); timetep is the analog step size; ha is hectare; site_area is the pel size.

On the other hand, the coupling simulation method for the high-resolution land utilization and forest landscape process is realized by the coupling simulation system for the high-resolution land utilization and forest landscape process, and comprises the following steps of:

Step 1, selecting historical land utilization data, temperature, rainfall, wind speed, solar radiation quantity, gradient and slope direction data, road and population space distribution map, land and city overall planning data and domestic production total value under different greenhouse gas emission scenes in the current and future as initial input data to be input into a land utilization change module;

step 2, calculating land utilization requirements, namely calculating total amount of various land utilization types in the future under different social and economic requirements and natural conditions based on land utilization data of an initial year by adopting a Markov Chain (Markov Chain), wherein a calculation formula is as follows:

S _t+1 ＝P _ij ×S _t

wherein: s is S _t And S is _t+1 Respectively representing the states of the land at the time t and the time t+1; p (P) _ij Representing that the land utilization type P is transformed into a matrix at the moment t;

step 3, calculating land utilization change probability; the land utilization change module utilizes an ANN neural network model to simulate the change probability and the spatial distribution of different land utilization types in the future based on socioeconomic and natural driving factor data;

the ANN neural network model consists of an input layer, a hidden layer and an output layer, wherein each neuron corresponds to a land utilization change driving factor respectively, and the specific formula is as follows:

G＝∑ _c w _1,k ×sigmoid(net _c (a,t))

Wherein: g represents the probability that the pixel a is converted into k in the land utilization type at the moment t; w (w) _1,k Is the adaptive weight between the hidden layer and the output layer; sigmoid (net) _c (a, t)) is the association function of the hidden layer with the output layer; net for writing _c (a, t) represents the signal sent by the pixel a on the first input layer to the neuron c at the time t, namely the intensity of the change of the pixel a in the type 1 land use type in the time t; w (w) _1,c And w _1,k Is an adaptive weight, which is distinguished by w _1,c Representing an adaptive weight relationship between the input layer and the hidden layer; x is x _i (a, t) is a function of t time variable 1 in relation to pel a in the input layer neuron;

step 4, calculating the neighborhood effect, namely the expansion strength of the land use types, wherein the expansion capacity of each land use type is stronger when the threshold value ranges from 0 to 1 and the value is close to 1; based on the land use status data of different periods, calculating the expansion strength of each land type according to the historical change trend of different land types, wherein the calculation formula is as follows:

NP＝n _b

wherein: NP is the number of plaques for all land types in a region; n is n _b Indicating the plaque amount of land utilization type b in a certain area; TA refers to the total area of all land utilization types in the area; land (land) ₂ Representing the total area of the 2 nd land use type in the area; area_am is the weighted average AREA of all plaques in the AREA; x is x _L2 A weight value representing the 2 nd land use type; neighbor represents the expansion strength of the land use type;

the self-adaptive inertial competition mechanism is a process of enabling an output result to continuously approach a target value through loop iteration, and an iteration loop formula is as follows:

wherein:representing the integrated probability of a pixel p transitioning from an initial land utilization type to a land type k at time t; omega shape _p,k ^t Representing the probability that land use type k appears at pel p; />The inertia coefficient of land use type k at time t is represented; sc _c→k Representing a conversion cost from land use type c to land use type k; />Representing the total number of pixels occupied by the earth type k at time t-1 under an NxN mole window; w (w) _k Is a variable weight between different land types; n is a molar neighborhood value in the land utilization change module; d (D) ^t-1 _k Indicating the difference between macroscopic demand and distribution of land type k at time t-1.

Step 5, utilizing a self-adaptive inertial competition mechanism based on roulette selection to combine neighborhood action and conversion rules, and realizing reasonable configuration of total pixel volume spatial distribution of each land type in the future based on the change probability distribution of different land types, thereby realizing simulation of land use change;

Step 6: the method comprises the steps of calculating the aboveground biomass by simulating land utilization changes, and quantifying the influence of climate change and land utilization change on the aboveground biomass of the forest by comparing the variation of the aboveground biomass under different climate change scenes and land utilization change patterns.

The simulated land utilization change is based on the total amount of each land utilization type in the future under different social and economic requirements and natural conditions, the number of each type of land is spatially distributed according to the probability, the neighborhood transformation rule and the limiting condition of each land utilization type in each grid unit, and the specific calculation formula is as follows:

L _predict ＝f(G,Neighbor,RE)→S _t+1

wherein: l (L) _predict A land utilization change simulation result at a certain moment; g is the probability that the land utilization type of the pixel a is converted into k at the moment t, and Neighbor represents the neighborhood conversion rule, namely the expansion strength of the land utilization type; RE is a socioeconomic limitation; s is S _t+1 Respectively representing the state of the land at the time t+1;

the above-ground biomass is calculated according to the number and the age of the tree, and the calculation formula is as follows:

wherein: AGB is the simulation result of the above-ground biomass at a certain moment; SDI (serial digital interface) _species1 A stand density index representing tree species 1; biomasscoff _species1 Calculating parameters for the aboveground biomass of tree species 1;

Quantifying the influence of climate change and land use change on forest land biomass by comparing the change of land biomass under different climate change situations and land use change situations; the calculation formula is as follows:

wherein: climate is the effect of Climate change on forest land biomass; LAND represents the influence of LAND utilization change on forest LAND biomass; AGB (AGB) _rcp The biomass on the forest land is the biomass on the future forest under different greenhouse gas emission scenes; AGB (AGB) _lucc Forest overground biomass under a future land utilization change pattern; AGB (AGB) _current Representing the current forest floor biomass.

The beneficial effects of adopting above-mentioned technical scheme to produce lie in:

the invention provides a high-resolution land utilization and forest landscape process coupling simulation system and method, so that the influence of future climate change and land utilization change on the carbon reserves of a forest ecological system in China is more accurately estimated. Compared with the existing ecological system process model, the model has the advantages of fewer required parameters, more convenient acquisition and higher operation efficiency. Forest resource survey data can be directly used for parameterizing the model and verifying the simulation accuracy of the model. The model can simulate the dynamic change of the structure and the function of an ecological system after the human operation and the natural disturbance with the spatial resolution of 30 meters, and the simulation precision can reach more than 80 percent.

The method realizes the high-efficiency and fine simulation of the dynamic state of the homeland resources under different global conditions by developing a high-resolution land utilization and natural ecosystem coupling model, is rich and perfect for land utilization change and land ecosystem process simulation theory, and provides a new thought and a new idea for scientific development and utilization of the homeland resources and a trampling green development mode. The current land utilization change simulation and the natural ecological system dynamic simulation are mutually independent, and interaction and synergistic effects between the society and the natural ecological system are not clearly explained, so that great uncertainty exists in the current land utilization simulation on the expression deficiency of the ecological system part and the land utilization mode ecological effect evaluation. By coupling the ecological system landscape process model and the land utilization change model, the simulation and release of key ecological system structures and functional elements are helpful for evaluating the adverse effects of land utilization and interference on the natural ecological system in novel urban construction work, the sensitivity and vulnerability of the natural ecological system to global changes are deeply known, the reasonable planning development direction and configuration resources of the government are guided, and the ecological management policy for actively adapting and relieving adverse effects of climate change is scientifically formulated in long term. Meanwhile, the comparison result of the model and the domestic autonomous mode can be helpful to enrich the sixth-stage coupling mode comparison plan (CMIP 6) and can be further provided for scientists in various countries to use in researching the influence and countermeasures of global and regional climate change.

Drawings

FIG. 1 is a schematic diagram of a coupling model framework for land utilization and forest landscape processes according to an embodiment of the present invention;

FIG. 2 is a flow chart of a coupling model of a land utilization and forest landscape process according to an embodiment of the present invention;

FIG. 3 is a diagram showing the verification of the model results in an embodiment of the present invention;

FIG. 4 is a diagram showing land utilization pattern change in a study area of years 2000-2100 according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of forest land biomass in a research area under different land use types and climate change situations for the future 100 years;

FIG. 6 is a schematic representation of the relative effects of land use and climate change on forest land biomass in accordance with the present invention.

Detailed Description

The following describes the embodiments of the present invention in detail with reference to the drawings.

In one aspect, the present invention provides a high-resolution land use and forest landscape process coupling simulation system, as shown in fig. 1, including a land use change module and a forest succession module;

the land utilization change module is used for calculating land utilization pattern changes under different social economy and natural conditions in the future by combining an adaptive inertial competition mechanism on the basis of a neural network model. The module can effectively couple social and economic elements and natural environment elements, realizes the simulation of land utilization change by designing conversion matrixes among different land types and roulette mechanisms, and well solves the problems of complex local conversion and parameter determination in the traditional cellular automaton method. The module can effectively combine the 'top-down' system dynamics with the 'bottom-up' cellular automaton, so that the large-scale future land utilization change fine simulation is possible.

The forest succession module is a space visual landscape model for simulating forest succession on a large space-time scale, the model simulates the influence of processes such as forest succession, seed diffusion, forest fire, harvesting, wind fall, plant diseases and insect pests, combustible material treatment and the like on the structure and the function of a forest ecological system, and the influence of tree species update, growth and death and Lin Fendong state change on the forest ecological system are respectively simulated on tree species, forest stand and landscape scale; simulating the processes of updating, growing, dying and the like of single trees on the tree species scale; simulating the processes of species colonisation, self-sparsity, under-forest regeneration, maturation and the like on a stand scale; on the landscape scale, simulating forest ecosystem processes such as forest succession, seed propagation, competition among seeds, forest fire, harvesting, wind fall, plant diseases and insect pests, combustible treatment and the like by tracking the existence and disappearance of the sample tree species; it specifically comprises the processes of tree growth, tree propagation and seed propagation.

The tree growth is determined by the growth space. The growth space refers to an area capable of providing nutrients, water and light for plant growth. In a limited growth space, different tree species compete for more growth space, causing death of the tree species and colonization of seedlings. In the forest succession module, the growth space is calculated by a forest Stand Density Index (SDI), a breast Diameter (DBH) and a stump density (TPA), and the forest stand density index is calculated according to the following formula:

Wherein: TPA is the standing tree density, i.e., the total number of all living standing trees with a height of more than 5cm per unit area; DBH is the diameter of the chest, which means the diameter of the trunk of standing wood from the chest height of the ground surface.

The process of calculating the density index of the stand is as follows: (1) The number of tree species in each diameter grade is converted into the number of tree species of standard tree with chest diameter of 10 inches (25.4 cm) through standing tree density index; (2) The maximum standing tree density (maxsi) is the number of plants per hectare that can accommodate the standard tree at most; (3) the minimum growth space is obtained based on a standard tree. The minimum growth space (GSi _standard) required for each standard tree can be calculated according to the following formula.

Wherein: maxSDI is the number of plants that can accommodate a certain tree species at most per hectare.

In the forest succession module, the simulation of tree growth is realized through the age-DBH relationship of different ages. All species will stop growing or die only when they reach the life of the species during the simulation. The age increase of the tree is estimated according to the simulated step size, and after one simulation, all the ages of each tree species are increased by one step size, and all the trees enter the next age. The relationship between age and breast diameter is usually determined according to a unitary harvesting table of different tree species of forest investigation, and the specific formula is as follows:

Wherein: maxDBH is the average maximum chest diameter when a certain tree species reaches the life; age is the current age of the tree species; longevity refers to the average maximum lifetime that this tree species can reach; a and k are formula constants derived from forest survey data.

The tree propagation is that after the tree death caused by forest fire or harvesting and the like or the physiological damage caused by drought, the tree is propagated by means of root tillers and the like. Aspen is one of the main broad-leaf tree species in northeast China, and the updating path of aspen is mainly root tiller propagation. In the forest succession module, the potential germination number and the seedling colonisation number are obtained by estimating the number of trees at each age level, and the potential germination number PPS of the seedlings is calculated as follows:

PPS＝NDT×VP

wherein NDT refers to the dead plant number of a certain tree species, VP refers to the asexual reproduction probability of the tree species

The tree propagation process in the forest succession module is as follows: (1) Determining whether the tree species is propagated by checking whether the tree species reaches a minimum propagation age; (2) Whether germinated seedlings can be colonised on a particular pel is determined by comparing the random germination probability (value range 0-1) with the tree species colonisation coefficient; (3) Modeling potential germination and seedling colonization numbers was achieved by tracking the number of tree plants at each age.

The seed propagation is a spatial process, affected by tree species properties and site conditions. The tree species attributes mainly include seed yield, seed effective propagation distance, maximum propagation distance, vegetative propagation ability and negative resistance of the plant of the parent tree. The site conditions include the relative density of site stand, the probability of tree species colonisation, the stand succession stage, etc. Seed sources can affect their surrounding site forest succession through seed propagation over a range of propagation distances. The seed propagation probability decreases exponentially with distance from the parent tree. The seed propagation distance is determined by the seed properties and propagation mode. The forest succession module defines a seed effective propagation distance (ED) within which the seed propagation probability is 95% and a seed maximum propagation distance (MD). The propagation probability of the seeds is only 5% beyond the maximum propagation distance. Between the two, the seed propagation probability P is derived from the following formula:

P＝e ^-b (x/MD)ED<x<MD

wherein MD is the maximum propagation distance of a certain tree species; x refers to the distance from the seed source; b is a seed propagation parameter; ED is the effective propagation distance of the seeds of the tree species; e refers to an exponential function based on a natural constant e.

Only a portion of the seeds may germinate under limited growth space conditions. Therefore, the number of potential seeds of a tree species is limited by the available growth space. The high variability of seed germination is caused by the maximum number of seeds produced by the mature parent tree and the variation in the maximum effective seed propagation distance for different tree species. The NPS formula for calculating potential seed number using growth space is as follows:

Wherein n refers to the number of simulated tree species; GSN represents the growth space required by germination of the available seed number of a certain tree species; GSO refers to currently occupying growth space; DBH refers to the chest diameter of the minimum age level of the tree species; cellsize refers to the pixel size. MaxSDI represents the number of plants per hectare that can accommodate a certain tree species at most.

In the forest succession module it is assumed that seed germination and colonisation are simultaneous. Uncertainty of seed sources and climate change can obviously influence the colonisation probability, and the colonisation probability is also influenced by environmental conditions such as sunlight, moisture, soil and the like. In this module, the environmental conditions can be determined by tree planting coefficients (SEC) on different land types. Tree species colonisation factors vary with space-time, reflecting landscape heterogeneity in terms of spatial and environmental variation. Thus, under different environmental conditions, the number of seedlings (NES) for a certain species of tree can be calculated as follows:

NES＝NPS×SEC _ecoregion

where NPS is the number of potential seeds of a tree species on a certain land type; SEC (SEC) _ecoregion And (3) establishing a group coefficient for the tree species on a certain land type.

The self-sparsity is a process that different tree individuals compete for light, water, heat and nutrition conditions, the difference among the individuals is larger and larger along with the growth, and the inferior tree individuals gradually die, namely the individual density gradually decreases along with the succession of forests. The stand density decreases with increasing tree size, either in homopolar or heteropolar forests. In the forest succession module, the self-sparsity phase starts to be entered when the site reaches or exceeds the self-sparsity condition. 3/2 self-sparse lines and occupied Growth Space (GSO) are often used to simulate the sparse process in forest distribution, the specific formulas are as follows:

In the formula, DBH _v And NT _V Mean chest diameter and plant number of a certain tree species at the V-th diameter level; 10inch refers to a standard tree with a chest diameter of 10 inches (25.4 cm); timetep is the analog step size; ha is hectare; site_area refers to the pixel size.

Different stages of stand development are defined according to 4 GSO thresholds: GSO (GSO) ₁ Representing the stage of occupying part of the growth space, all the reached species may sprout and colonise, except the most negative species; GSO (GSO) ₁ =1, canopy-closed stage, when only the most yin-tolerant tree species germinate and colonise; GSO (GSO) ₃ Tree species already occupy the whole growth space, and the vigorous competition leads to the inability of all tree species to colonise; GSO (GSO) _thinning From the sparsity stage, the stand density begins to decrease.

On the other hand, the coupling simulation method for the high-resolution land utilization and forest landscape process is realized by the coupling simulation system for the high-resolution land utilization and forest landscape process, as shown in fig. 2, and comprises the following steps:

step 1, selecting historical land utilization data, temperature, rainfall, wind speed and solar radiation quantity, gradient and slope direction data, road and population space distribution map, land and city overall planning data and domestic production total value socioeconomic data under different greenhouse gas emission scenes in the current and future as initial input data to be input into a land utilization change module;

Step 2, calculating land utilization requirements; the total amount of various land utilization types in the future under different social and economic requirements and natural conditions is calculated based on land utilization data of the initial year by adopting a Markov Chain (Markov Chain), and the calculation formula is as follows:

S _t+1 ＝P _ij ×S _t

G＝∑ _c w _1,k ×sigmoid(net _c (a,t))

wherein: g represents the probability that the pixel a is converted into k in the land utilization type at the moment t; w (w) _1,k Is the adaptive weight between the hidden layer and the output layer; sigmoid (net) _c (a, t)) is the association function of the hidden layer with the output layer; net for writing _c (a, t) represents the signal sent by the pixel a on the first input layer to the neuron c at the time t, namely the intensity of the change of the pixel a in the type 1 land use type in the time t; w (w) _1,c And w _1,k Is an adaptive weight, which is distinguished by w _1,c Representing inputAn adaptive weight relationship between the layer and the hidden layer; x is x _i (a, t) is a function of t time variable 1 in relation to pel a in the input layer neuron.

NP＝n _b

wherein: NP is the number of plaques for all land types in a region; n is n _b Indicating the plaque amount of land utilization type b in a certain area; TA refers to the total area of all land utilization types in the area; land (land) ₂ Representing the total area of the 2 nd land use type in the area; area_am is the weighted average AREA of all plaques in the AREA; x is x _L2 A weight value representing the 2 nd land use type; neighbor represents the expansion strength of the land use type.

Step 5, utilizing a self-adaptive inertial competition mechanism based on roulette selection to combine neighborhood action and conversion rules, and realizing reasonable configuration of total pixel volume spatial distribution of each land type in the future based on the change probability distribution of different land types, thereby realizing simulation of land use change; the self-adaptive inertial competition mechanism is a process of enabling an output result to continuously approach a target value through loop iteration, and an iteration loop formula is as follows:

Wherein:representing the integrated probability of a pixel p transitioning from an initial land utilization type to a land type k at time t; omega shape _p,k ^t Representing the probability that land use type k appears at pel p; inertia ^t _k The inertia coefficient of land use type k at time t is represented; sc _c→k Representing a conversion cost from land use type c to land use type k; />Representing the total number of pixels occupied by the earth type k at time t-1 under an NxN mole window; w (w) _k Is a variable weight between different land types; n is a molar neighborhood value in the land utilization change module; d (D) ^t-1 _k Indicating the difference between macroscopic demand and distribution of land type k at time t-1.

L _predict ＝f(G,Neighbor,RE)→S _t+1

The northeast forest of China accounts for 29.9 percent of the total area of the natural forest of China, and about 1/3 of the carbon reserves of the forest of China are stored. For over half a century, northeast forest areas have been used as one of the main wood production bases in China, forests have undergone long-time high-strength deforestation, and most forests are currently in early succession stages such as young forests or middle-aged forests. Long-term high-intensity harvesting significantly changes the tree species composition, structure and carbon reserves of forests in northeast areas. The main updated tree species of the forests after harvest are pioneer tree species such as white birch, larch, aspen, basswood and the like, and the stand age is basically between 40 and 70 years. Studies have shown that these early succession tree species in the climate change background may continue to reach life in the next decades, such that the mortality of the forest exceeds the growth rate, resulting in the risk of the northeast forest ecosystem being converted from carbon sink to carbon source. In addition, with the rapid development of the Chinese socioeconomic performance, a large number of forests are becoming non-forests. Comprehensively considering the influence of climate change and land utilization change on the forest ecosystem, the countermeasure for optimizing the structure and improving the function is provided, and the countermeasure is an urgent need for sustainable management of the northeast forest ecosystem in the future. The method is characterized in that the Changbai mountain of Jilin province is selected as a research area, the high-resolution land utilization and forest landscape process coupling model is utilized to simulate the future succession dynamic change of the northeast typical forest ecological system under the combined action of the climate change and the land utilization change, the influence of the climate change and the land utilization change on the overground biomass of the northeast typical forest ecological system in the future 100 years is quantified, and scientific basis is provided for formulating reasonable economic development and forestry management targets in the northeast forest area.

In order to quantify the influence of climate change and land use change on the northeast typical forest ecosystem of China, the present embodiment sets 4 climate change scenes such as current, RCP2.6, RCP4.5 and RCP8.5, and 2 land use change plans (taking land use change and land use out into consideration) for 8 simulation plans in total, as shown in Table 1. The existing weather undisturbed plan does not consider land utilization change, and represents potential aboveground biomass of forests under the existing weather conditions; the current climate land utilization plan considers the influence of land utilization change and represents the actual situation of the above-ground biomass under the current climate condition. RCP2.6, RCP4.5 and RCP8.5 climate change scenarios the influence of land use changes in the study area of the next 100 years was not considered, representing potential aboveground biomass of forests in the study area in future climate change scenarios. The land use plans in the climate change scenarios of RCP2.6, RCP4.5 and RCP8.5 respectively take the influence of climate change and land use change into consideration, and represent the actual condition of the biomass on the research area under the future climate change condition. With a 10 year time step, the change of the ground biomass of the next 100 years is simulated, and all simulation plans are repeated 5 times to reduce the uncertainty of the model.

Table 1 simulation plan settings

The scheme of the embodiment of the invention selects the data sources and processes the data as follows:

the data for the model of this embodiment mainly includes: 1) Historical land use current situation diagrams of 5 periods from 1990 to 2010; 2) Future prediction data mainly comprises various land grid percentage data, population and economic data under different RCPs scenes; 3) Natural element data, which is used for driving the natural environment elements of land utilization change to mainly comprise soil, climate and topography; 4) Socioeconomic data for driving a change in land use, mainly including data of population data, global value for domestic production (GDP) road network, and city center distance, etc.; 5) Forest phase diagrams and forest field investigation data in Changbai mountain areas.

To reflect the effect of climate change on forest carbon reserves, a current climate scenario (which simulates as a control scenario, temperature and precipitation conditions are kept at the current average level and do not change), an RCP2.6 scenario (climate warming trend is low), an RCP4.5 scenario (climate warming trend is strong), and an RCP8.5 scenario (climate warming trend is strongest) are selected. The RCP2.6, RCP4.5 and RCP8.5 scenario temperature, precipitation, wind speed and radiation data for CCSM4 mode prediction were selected from CMIP5 (Coupled Model Inter-comparison Project, CMIP 5) program (2007-2100). The CCSM4 mode (Community Climate System Model, CCSM 4) is one of the internationally new generation of coupled climate modes, and consists of five major components of atmosphere, sea, land, sea ice and couplers, and the simulation performance is greatly improved compared with the prior climate mode. Climate change data Source IPCC climate data website (https:// kgf-node. Llnl. Gov). For the current climate scenario data, 1970-2010 weather observation data (http:// data. Cma. Cn/site/index. Html) was selected. In order to obtain meteorological parameters of each land type, observation data of a research area and surrounding meteorological stations are selected, the influence of longitude and latitude, altitude and topography is considered, and the temperature, rainfall and wind speed data of the research area are spatially interpolated by adopting a regression Kriging method. Based on the lowest and highest daily temperature data, the average daily incident radiation of the study area is estimated by using a Campbell empirical model.

The land utilization change module is used for inputting a land utilization change driving factor by the ANN neural network module, and the following factors are selected according to actual conditions and the availability of data: (1) The terrain condition is one of key factors for determining the land utilization type, wherein the elevation and the gradient are main factors for determining the land utilization, and DEM data with 30m resolution is adopted; (2) Traffic conditions and town developments have an important attractive effect on land utilization, selecting distances to general highways, to district centers and to town centers as accessibility factors. The method is obtained through analysis of Euclidean distance tools of geographic information system software ArcGIS. Setting the training sampling proportion of the neural network, and selecting a random sampling mode to sample training samples of various lands, thereby realizing the training of the neural network. And finally calculating to obtain the suitability probability map of the land utilization type on each pixel by combining the distribution condition of each driving factor after the standardized treatment.

The cellular automaton module based on the self-adaptive inertia mechanism takes multi-class or double-class space land utilization data as initial input data, and targets of the change quantity of each land utilization type are required to be preset. Firstly, a Markov prediction model is adopted to obtain the preset quantity of the change of each land utilization type. And then determining the interconversion difficulty between different land types according to experience (0-1, 0 indicates that conversion is not allowed, 1 indicates that conversion can be freely performed), and finally setting a limit occurrence area for interconversion of land utilization types. In the setting of model parameters, the simulation iteration target frequency is set to be 500, the neighborhood size is set to be 3 multiplied by 3, and finally, the land utilization change simulation is realized.

The space parameters of the forest succession module are a site type diagram, a tree species composition diagram, a management area diagram, a Lin Huoji drawing and the like. The land type map is generated from the study area forest map, the 30m resolution digital elevation map, and the study area land use status map and road map (1:100000). And obtaining an initial tree species composition diagram by combining forest resource investigation data with a 2000 year study area forest phase diagram. The influence of human activities such as land utilization change on a forest ecosystem usually occurs in a transition zone of forests and artificial landscapes, and the spatial resolution of a model is set to be 100m multiplied by 100m under the condition of balancing the existing computer resources and describing the relationship between fine-scale ecological processes as far as possible.

The non-space parameters mainly comprise tree species life history attributes, tree species colonisation probability, forest fire parameters, harvest parameters and the like. The tree species life history attribute parameters are mainly obtained by referring to related documents, existing researches, field investigation and the like, as shown in table 2. The tree species colonisation factor (SEC) may reflect the effect of environmental changes on tree species, reflecting the direct effect of climate changes on forest ecosystems by tree species colonisation probabilities within different land types. The tree species colonisation probability under the current climate conditions is derived from the literature of the research area on forest landscape models. Referring to the existing research, the LINKAGES v3.0 model is utilized to estimate the tree species colonisation probability under different climate change situations.

Table 2 forest succession module tree species life history attributes

The simulation results of different climate change situations are compared to evaluate the influence of climate change and land utilization change on the biomass on the forest land of the Changbai mountain. Short term (0-20 years), medium term (30-50 years) and long term (60-100 years) are chosen to represent the changes of the spatial-temporal patterns of forest land biomass reflected in the research area. The "releimpo" software package in the R statistical software was used to estimate the relative impact of climate change and land use change on forest land biomass in the study area. The Relaimpo software package uses a correlation weighting method (pmvd) to calculate average weights for estimating the relative importance of the predicted variables in the multiple linear regression. The study was conducted by this method to estimate the relative impact of climate change and land use change on aboveground biomass in 3 phases of short, medium and long term.

Verifying a model result;

the embodiment adopts a data segmentation method (data splitting approach) to verify the model result. Firstly, selecting 70% of forest class-II investigation data and a 2000 year forest map data parameterized model, and comparing differences between a 2000 year simulation result and 30% of class-II investigation data through variance analysis, wherein a comparison variable is aboveground biomass; if the difference is obvious (unacceptable), the model initialization parameters (parameters such as a growth curve, mortality, seed quantity and the like) are adjusted, and the model is re-simulated; the verification process is repeated until the discrepancy is acceptable. Next, the model succession module is run for 10 years (to 2010), and the 2010 forest survey data is selected to verify the model simulation results

Model initialization values and overground biomass simulation results in 2010 are close to contemporaneous forest survey data. Comparison of the initialized aboveground biomass with 384 forest survey sample data in 2000 shows that the initialized aboveground carbon reserves are close to the 2000 forest survey data, and the correlation between the initialized aboveground biomass and the 384 forest survey sample data is 0.85. The ground biomass simulation result in 2010 is close to the data of the 407 forest survey samples in 2010, and has high correlation and R ² =0.85, as shown in fig. 3.

Analysis of results

(1) Land use change analysis:

land utilization changes in Changbai mountain areas mainly comprise occupation of forests by construction lands such as cities, rural residents and roads due to social and economic development, conversion of forests into farmlands by deforestation, conversion of forests into grasslands required by animal husbandry development and the like. At the same time there is also a return to forests and other land changes to forests due to forests and national policies. The land utilization and forest landscape process coupling model integrates driving factors such as nature, society and economy, combines the land utilization current situation to simulate the suitability distribution probability of each land type under a preset scene, introduces a self-adaptive inertial competition mechanism based on roulette selection, and is used for processing uncertainty and relative complexity of the change of various land types under the synergistic effect of nature, society and economy, thereby realizing high-precision land utilization change simulation. Model results show that the farmland of the research area occupies 27692ha forest land in the future 100 years, and is mainly distributed in the south and southwest of the research area. 3472ha woodland was converted to construction land in 2000-2100. As shown in fig. 4, the construction land expansion pattern of the study area shows that town development is gradually expanding around the current town as a center. At the same time, 478 and 1049ha farms and grasslands, respectively, are restored to woodlands, as shown in table 3, under the influence of the national ecological protection policy.

Table 3 land use change units for study area 2000-2100 years: ha (ha)

(2) Influence of climate and land use changes on aboveground biomass

Under different climate change scenarios, the above-ground biomass in the research area gradually increases with the simulation time, but the land utilization change will reduce the above-ground biomass accumulation speed, as shown in fig. 5. Compared with the existing climate, the on-ground carbon reserves of the RCP4.5 and RCP8.5 climate change scenes do not consider the land utilization change plan are greatly increased and are respectively 7% and 12% higher than those of the existing climate which does not consider the land utilization change plan. The above-ground biomass of the climate change scenario of RCP2.6, RCP4.5 and RCP8.5 considered land use plans is 216.5+ -35.1, 222.2+ -38.0 and 226.7 + -40.0 t/ha, respectively, significantly greater than that of the current climate considered land use plans throughout the simulation phase (2000-2100 years). The RCP8.5 climate change scenario considers that the land use plan has an above ground biomass significantly greater than RCP2.6 and the current climate consideration land use plan has an above ground biomass 4% and 5% greater than RCP2.6 and the current climate scenario plan, respectively.

Climate change and land utilization have a large impact on the forest land biomass of the Changbai mountain as shown in FIG. 6. Climate change significantly increases the aboveground biomass of the forest of the Changbai mountain, and the effect of climate change on aboveground biomass in the last 30 years of simulation is higher than that in 40 years after simulation. The effect of climate change on the above-ground biomass is greater than the effect of land utilization at all 3 different stages. The relative impact of land use changes on the land biomass in the study area was 27% and 32% in the short-term and mid-term, respectively. In addition, the influence of climate change on the above-ground biomass tends to decrease in 3 different stages, while the influence of land use change on the above-ground biomass tends to increase.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the corresponding technical solutions, which are defined by the scope of the appended claims.

Claims

1. A high-resolution land utilization and forest landscape process coupling simulation system is characterized in that: the system comprises a land utilization change module and a forest succession module;

The growth of the tree is determined by a growth space, and the growth space refers to an area capable of providing nutrients, water and illumination for plant growth; in a limited growth space, different tree species compete for more growth space, thereby causing death of tree species and colonization of seedlings; in the forest succession module, the growth space is calculated by a forest stand density index, a breast diameter and a standing tree density, wherein the formula of the forest stand density index is as follows:

PPS＝NDT×VP

P＝e ^-b (x/MD)ED<x<MD

the self-sparsity is a process that different tree individuals compete for light, water, heat and nutrition conditions, the difference among the individuals is larger and larger along with the growth, and the inferior tree individuals die, namely the individual density is gradually reduced along with the succession of forests; the 3/2 self-sparse lines and the occupied growth space are used to simulate the self-sparse process in the forest distributing and spreading process, and the specific formula is as follows:

in the formula, DBH _v And NT _V Mean chest diameter and plant number of a certain tree species at the V-th diameter level; 10inch refers to a standard tree with a chest diameter of 10 inches; longevity refers to the average maximum life that this tree species can reach, and timetep is the simulated step size; ha is hectare; site_area is the pel size.

2. The high resolution land use and forest landscape process coupling simulation system of claim 1, wherein the 4 GSO thresholds comprise: GSO (GSO) ₁ Occupying part of the growth space stage, and all the reached tree species are germinated and colonised except the most negative-tolerant tree species; GSO (GSO) ₁ =1, canopy-closed stage, when only the most yin-tolerant tree species germinate and colonise; GSO (GSO) ₃ Tree species already occupy the whole growth space, and the vigorous competition leads to the inability of all tree species to colonise; GSO (GSO) _thinning From the sparsity stage, the stand density begins to decrease.

3. The high-resolution land utilization and forest landscape process coupling simulation method is realized by the high-resolution land utilization and forest landscape process coupling simulation system as claimed in claim 1, and is characterized by comprising the following steps:

step 2, calculating land utilization requirements, namely calculating total amount of various land utilization types in the future under different social and economic requirements and natural conditions based on land utilization data of an initial year by adopting a Markov chain, wherein a calculation formula is as follows:

S _t+1 ＝P _ij ×S _t

G＝∑ _c w _1,k ×sigmoid(net _c (a,t))

NP＝n _b

Wherein: NP is the number of plaques for all land types in a region; n is n _b Indicating the plaque amount of land utilization type b in a certain area; TA refers to the total area of all land utilization types in the area; land (land) ₂ Representing the total area of the 2 nd land use type in the area; AREA (AREA)AM is the weighted average area of all plaques in the area; x is x _L2 A weight value representing the 2 nd land use type; neighbor represents the expansion strength of the land use type;

4. The coupling simulation method for high-resolution land utilization and forest landscape process according to claim 3, wherein the adaptive inertial competition mechanism in step 4 is a process of continuously approaching the output result to the target value through loop iteration, and the iterative loop formula is as follows:

Wherein:representing the integrated probability of a pixel p transitioning from an initial land utilization type to a land type k at time t; omega shape _p,k ^t Representing the outline of the land use type k appearing at pixel pA rate; inertia ^t _k The inertia coefficient of land use type k at time t is represented; sc _c→k Representing a conversion cost from land use type c to land use type k; />Representing the total number of pixels occupied by the earth type k at time t-1 under an NxN mole window; w (w) _k Is a variable weight between different land types; n is a molar neighborhood value in the land utilization change module; d (D) ^t-1 _k Indicating the difference between macroscopic demand and distribution of land type k at time t-1.

5. The coupling simulation method for high-resolution land utilization and forest landscape process according to claim 3, wherein the simulated land utilization change in step 6 is based on different socioeconomic requirements and total amount of future land utilization types under natural conditions, and the number of land utilization types is spatially distributed according to probability, neighborhood transformation rules and constraint conditions of each land utilization type in each grid unit, and the specific calculation formula is as follows:

L _predict ＝f(G,Neighbor,RE)→S _t+1

wherein: AGB is the simulation result of the above-ground biomass at a certain moment; SDI (serial digital interface) _species1 A stand density index representing tree species 1; biomasscoff _species1 Is the land of tree species 1Calculating parameters by using the upper biomass;