CN117787503A - Estimation and prediction method for dead area of grassland in permafrost region of Qinghai-Tibet plateau - Google Patents
Estimation and prediction method for dead area of grassland in permafrost region of Qinghai-Tibet plateau Download PDFInfo
- Publication number
- CN117787503A CN117787503A CN202410088851.XA CN202410088851A CN117787503A CN 117787503 A CN117787503 A CN 117787503A CN 202410088851 A CN202410088851 A CN 202410088851A CN 117787503 A CN117787503 A CN 117787503A
- Authority
- CN
- China
- Prior art keywords
- species
- qinghai
- vegetation
- survival
- tibet plateau
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 22
- 230000008859 change Effects 0.000 claims abstract description 19
- 230000007613 environmental effect Effects 0.000 claims abstract description 12
- 230000003068 static effect Effects 0.000 claims abstract description 7
- 230000004083 survival effect Effects 0.000 claims description 109
- 230000007423 decrease Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 230000035899 viability Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 abstract description 22
- 241000894007 species Species 0.000 description 77
- 241000196324 Embryophyta Species 0.000 description 14
- 238000009826 distribution Methods 0.000 description 13
- 230000000694 effects Effects 0.000 description 12
- 239000002689 soil Substances 0.000 description 8
- 238000011835 investigation Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 230000009746 freeze damage Effects 0.000 description 5
- 230000029553 photosynthesis Effects 0.000 description 5
- 238000010672 photosynthesis Methods 0.000 description 5
- 239000010410 layer Substances 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 230000007774 longterm Effects 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 230000002123 temporal effect Effects 0.000 description 3
- 230000005068 transpiration Effects 0.000 description 3
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 2
- 241001465754 Metazoa Species 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 229910052799 carbon Inorganic materials 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000001932 seasonal effect Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000010206 sensitivity analysis Methods 0.000 description 2
- 238000009827 uniform distribution Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 229920002430 Fibre-reinforced plastic Polymers 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000005842 biochemical reaction Methods 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 230000021523 carboxylation Effects 0.000 description 1
- 238000006473 carboxylation reaction Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 239000007795 chemical reaction product Substances 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000006735 deficit Effects 0.000 description 1
- 230000007850 degeneration Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000011151 fibre-reinforced plastic Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 239000011229 interlayer Substances 0.000 description 1
- 230000000116 mitigating effect Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 235000015097 nutrients Nutrition 0.000 description 1
- 230000005868 ontogenesis Effects 0.000 description 1
- 230000008288 physiological mechanism Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a method for estimating and predicting dead areas of grasslands in permafrost regions of Qinghai-Tibet plateau, which comprises the following steps: s1, data arrangement is carried out on environmental data of permafrost areas of Qinghai-Tibet plateau and vegetation species of grasslands to form a data set; s2, establishing a static basic model of a year-year dead area estimation and prediction model of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau by using a Logistic function; s3, estimating dead area of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau and assigning a time change inflection point performance parameter of a prediction model; s4, objectively evaluating experience of a vegetation dead area estimation and prediction model in a permafrost region of the Qinghai-Tibet plateau; s5, estimating the dead area of grassland vegetation in the permafrost region of the Qinghai-Tibet plateau and updating the parameters of the prediction model. The method considers the parameter differences in the fitting of the vegetation dead functions in the plot, time (year), species and species, and improves the accuracy and reliability of the prediction of the occurrence probability of the vegetation dead disasters of the grasslands.
Description
Technical Field
The invention relates to the technical field of natural disaster prediction, in particular to a method for estimating and predicting dead areas of grasslands in permafrost regions of Qinghai-Tibet plateau.
Background
The perennial frozen soil region grassland of the Qinghai-Tibet plateau occupies important positions in the global environment and the ecological system, is not only an important water source conservation region of Asia, but also habitat of numerous rare wild animals and plants, and has irreplaceable effects on maintaining biodiversity and ecological balance. These grasslands possess rich ecosystem service functions including carbon storage, climate regulation, soil and water conservation, etc., and have profound effects on global climate change mitigation and adaptation. Meanwhile, the grasslands in the permafrost areas of the Qinghai-Tibet plateau are also vital to the life of local residents, the animal husbandry is a main economic source of local herding, and the health condition of the grasslands is directly related to the life of local people and the sustainable development of social economy. However, as global climate warms, permafrost degenerates, causing grassland degeneration and the grassland ecosystem of the Qinghai-Tibet plateau is facing serious problems such as increased dead area. Therefore, by accurately estimating the dead area of the grassland, the method has important significance for better understanding and monitoring the health condition of the grassland ecological system and timely finding the trend and the reason of ecological degradation.
However, existing investigation techniques are time-consuming and labor-consuming in acquiring data, particularly because of the harsh geographical environment of the Qinghai-Tibet plateau, inconvenient traffic, large-scale field investigation all require considerable investment of capital and human resources, and some field estimation methods may not be capable of achieving high accuracy, particularly in complex and varying natural environments. Meanwhile, the dynamic and complex of the grassland ecological system makes accurate estimation of dead areas more difficult. The method relying on remote sensing may be limited by the time resolution and spatial resolution of the satellite, cloud cover, image quality, and other factors, so that the change in a smaller scale or a short time scale cannot be effectively monitored. In addition, due to the fact that real-time conditions of grassland dead spots can be affected by seasonal changes, climate anomalies and the like, the existing estimation technology can not comprehensively consider the factors, and therefore great challenges are brought to accurate estimation and prediction of grassland dead spots in permafrost areas of Qinghai-Tibet plateau.
Disclosure of Invention
The invention aims to provide an accurate and reliable estimation and prediction method for the dead area of a grassland in a permafrost region of a Qinghai-Tibet plateau.
In order to solve the problems, the method for estimating and predicting the dead area of the grassland in the permafrost region of the Qinghai-Tibet plateau comprises the following steps:
s1, data arrangement is carried out on environmental data of permafrost areas of Qinghai-Tibet plateau and vegetation species of grasslands to form a data set;
s2, establishing a static basic model of a year-year dead area estimation and prediction model of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau by using a Logistic function, wherein the model is as follows:
K～normal(2,1)
wherein:the method comprises the steps of (1) circularly processing a year chain, namely, finishing and forming all years in a time interval of environmental data of a permafrost region of a Qinghai-Tibet plateau and a grassland vegetation species data set in the step of S1; k (K) lat Represents the potential survival rate per year under ideal conditions; />An inverse Logistic function representing annual survival rate; k (K) μ Normal (2, 1) represents an a priori function showing average survival; k (K) μ The intrinsic survival rate of a specific species, namely the survival rate which is free from external environmental factors and time; />Represents the p-like average survival rate of the ith species; />Represents the average survival rate of the ith species in the x year; epsilon i Representing an ith species viability fit residual; θ represents the grassland vegetation survival probability; DBH (DBH) i Represents the average height of adult plants of the i-th species; threshold represents a fitting threshold demarcation point; k is the survival probability of most of the individual's adulthood; r1 is the rate at which the survival rate of the S-shaped portion increases to reach an arbitrary size point at which the survival probability becomes K; p1 is the size corresponding to the inflection point of the functional form of the S-shaped portion; r2 is the rate of decline in annual survival rate of species of the inverted S-shaped portion approaching its maximum size; p2 is the inflection point or threshold height or size corresponding to 50% of the annual survival probability of the inverted S-shaped portion;
s3, estimating dead area of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau and assigning a time change inflection point performance parameter of a prediction model:
the average survival rate after adjustment according to the sample side information is expressed as a priori:
K P ～Normal(0，σ P )；σ P is p-like standard difference;
the average survival rate after adjustment according to the time information is a priori expressed as:
K_T～Normal(0，σ T )；σ T standard deviation of time T;
the residual after adjustment according to the existing information is expressed as:
∈～Normal(0，σ ∈ )；σ ∈ standard deviation for residual e;
the a priori representation of the standard error coefficient is:
σ P ，σ T ，σ ∈ ～Gamma(3，4)；
the a priori representation of the p1 survival parameters is:
P1～Normal(min(dbh)，1)；
s4, objectively evaluating experience of a vegetation dead area estimation and prediction model in a permafrost region of the Qinghai-Tibet plateau;
the a priori representation of the p2 survival parameters is:
P2～Normal(min(dbh)，1)；
the a priori representation of the r1 survival parameters is:
r1～Uniform(0，2)；
the a priori representation of the r2 survival parameters is:
r2～Uniform(-2，0)；
s5, estimating the dead area of grassland vegetation in the permafrost region of the Qinghai-Tibet plateau and updating the parameters of the prediction model.
Compared with the prior art, the invention has the following advantages:
1. based on the characteristic attribute of dead areas of grassland vegetation in the permafrost region of the Qinghai-Tibet plateau, the invention improves the accuracy and reliability of the prediction of the occurrence probability of dead disasters of the grassland vegetation by independently fitting a survival function for each species in a multi-layer model frame, taking into consideration the parameter differences during fitting of dead functions of vegetation in the sample land, time (year), species and species.
2. According to the invention, through coupling 5 basic description grassland vegetation survival parameters, an S-shaped and reverse S-shaped survival probability Logistic function is formed, and the relation between mortality and individual development is focused through plant size, so that the reasons of individual death of different plants or the threshold difference of environmental parameters are realized.
3. The invention solves the problem of conditional probability distortion and improves the accuracy of model prediction by introducing a No-U-Turn-Sampler (NUTS; 5,000 iterations on three chains) method.
Drawings
The following describes the embodiments of the present invention in further detail with reference to the drawings.
FIG. 1 is a flow chart of the present invention.
Fig. 2 is a graph of results of estimation and prediction model of dead area of grassland vegetation in perennial frozen soil areas of Qinghai-Tibet plateau based on a bayesian interlayer structure according to an embodiment of the present invention.
Detailed Description
As shown in fig. 1, the method for estimating and predicting the dead area of the grassland in the permafrost region of the Qinghai-Tibet plateau comprises the following steps:
s1, data arrangement is carried out on environmental data of permafrost areas of Tibet plateau and grassland vegetation species to form a data set.
S2, establishing a static basic model of a year-year dead area estimation and prediction model of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau by using a Logistic function, wherein the model is as follows:
K～normal(2,1)
wherein:the method comprises the steps of (1) circularly processing a year chain, namely, finishing and forming all years in a time interval of environmental data of a permafrost region of a Qinghai-Tibet plateau and a grassland vegetation species data set in the step of S1; k (K) lat Represents the potential survival rate per year under ideal conditions; />An inverse Logistic function representing annual survival rate; k (K) μ Normal (2, 1) represents an a priori function showing average survival; k (K) μ The intrinsic survival rate of a specific species, namely the survival rate which is free from external environmental factors and time; />Represents the p-like average survival rate of the ith species; />Represents the average survival rate of the ith species in the x year; epsilon i Representing an ith species viability fit residual; θ represents the grassland vegetation survival probability; DBH (DBH) i Represents the average height of adult plants of the i-th species; threshold represents a fitting threshold demarcation point; k is survival of most of the adult life of an individualProbability; r1 is the rate at which the survival rate of the S-shaped portion increases to reach an arbitrary size point at which the survival probability becomes K; p1 is the size corresponding to the inflection point of the functional form of the S-shaped portion; r2 is the rate of decline in annual survival rate of species of the inverted S-shaped portion approaching its maximum size; p2 is the corresponding inflection point or threshold height or size for 50% of the annual survival probability of the inverted S-shaped portion.
S3, estimating dead area of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau and assigning a time change inflection point performance parameter of a prediction model:
the average survival rate after adjustment according to the sample side information is expressed as a priori:
K P ～Normal(0，σ P )；σ P is p-like standard difference; namely: the survival rate of p-plot can be 0 or the average value is 0, the standard deviation is sigma P Is fitted to the normal distribution of (c).
The average survival rate after adjustment according to the time information is a priori expressed as:
K_T～Normal(0，σ T )；σ T standard deviation of time T; namely: the survival rate at time T can be set to 0 by a desired value or 0 by an average value, and the standard deviation is sigma T Is fitted to the normal distribution of (c).
The residual after adjustment according to the existing information is expressed as:
∈～Normal(0，σ ∈ )；σ ∈ standard deviation for residual e; namely: the residual error E satisfies a desired value of 0 or an average value of 0, and the standard deviation is sigma ∈ Is distributed in the normal direction.
The a priori representation of the standard error coefficient is:
σ P ，σ T ，σ ∈ gamma (3, 4); namely: sigma (sigma) P ，σ T ，σ ∈ A gamma distribution of 3 shape and 4 scale is satisfied.
The a priori representation of the p1 survival parameters is:
p1 to Normal (min (dbh), 1); namely: p1 is an n-too distribution with a desired value of minimum adult plant height and standard deviation of 1.
S4, objectively evaluating experience of a vegetation dead area estimation and prediction model in a permafrost region of the Qinghai-Tibet plateau;
the a priori representation of the p2 survival parameters is:
p2 to Normal (max (dbh), 1); namely: p2 is an n-too distribution with a desired value of minimum adult plant height and standard deviation of 1.
The a priori representation of the r1 survival parameters is:
r1 to form (0, 2); namely: r1 is a uniform distribution with a minimum of 0 and a maximum of 2.
The a priori representation of the r2 survival parameters is:
r 2-form (-2, 0); namely: r2 is a uniform distribution with a minimum of-2 and a maximum of 0.
S5, estimating the dead area of grassland vegetation in the permafrost region of the Qinghai-Tibet plateau and updating the parameters of the prediction model.
Examples
As shown in fig. 1-2, the method for estimating and predicting the dead area of the grassland in the permafrost region of the Qinghai-Tibet plateau comprises the following steps:
s1, data arrangement is carried out on environmental data of permafrost areas of Qinghai-Tibet plateau and vegetation species of grasslands to form a data set;
the system sorts up 110,551 observation records from 1975 to 2023 to form a grassland vegetation wither data set in the Qinghai-Tibet plateau permafrost region, which comprises 13,513 independent sample parties, covers 135 species, 54 genera and 79 families, and relates to three grassland vegetation types: high-cold meadows, high-cold meadows and swamps. These plots were distributed along a soil moisture gradient, with soil moisture content ranging between 5% and 40%. The plot was investigated on average 12 times, 20 long-term plots were re-measured every 2 years over the first 10 years, followed by measurements at 3-4 year intervals, recording species abundance, coverage and dead area.
The present invention uses criteria of at least 400 observations per species to model functional individual size dependence of species survival to ensure adequate statistical power. A dataset was generated containing 81 species, 8,314 individual plant individuals, 74,135 observations and 492 death events. Single species were distributed between 1 and 16 plots, evenly distributed over 6.1 plots, single species were investigated 2 to 17 times (10.4 times average), spanning a period of 2 to 44 years (13 years average).
S2, establishing a static basic model of a year-year dead area estimation and prediction model of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau;
because the direct dead rate is not well quantitatively calculated, the present invention indirectly reflects the vegetation dead rate, i.e., vegetation dead rate = 1-vegetation survival (a potential or unobserved variable), with an estimate of the annual grassland vegetation survival probability, whereas vegetation survival can be achieved by fitting a survival function for each species alone in a multi-layered bayesian model framework (equation (1)), with "plot" and "year" as the varying effects (also commonly referred to as "random effects"), i.e., likelihood equations are a Bernoulli distribution:
Surivival i，k，t bernoulli (θ) … equation (1)
Surivval in equation (1) i，k，t The survival rate of the observation object i at time t in the kth plot is represented.
Estimation of annual grassland vegetation survival probability (a potential or unobserved variable) is achieved by fitting a survival function for each species individually in a multi-layer model framework, with "plot" and "year" as the changing effects (also commonly referred to as "random effects"). The first step is therefore to estimate plant size dependent grassland survival by fitting a functional form individually to each species in multiple surveys and grasslands. The model used to describe the survival of grassland vegetation combines five survival parameters to form a sigmoid and inverted sigmoid survival probability logistic function. These parameters correspond to the lowest survival rate per year (K; survival probability of the individual in adulthood most of the time), the rate of increase of survival to a rate (r 1) reaching the point of arbitrary size at which the survival probability becomes K, the size (p 1) corresponding to the inflection point of the functional form of the first S-shaped portion, and the respective r2 and p2 parameters of the functional form of the inverted S-shaped portion, respectively correspond to the rate of decrease of the survival rate per year of the species approaching its maximum size, and the size of the inflection point (species individual height) corresponding to the annual survival probability of 50%.
In the present invention, individual size scale (K, survival of most of the plant in adulthood) after logistic transformation was simulated separately for each plant species as a function of species specific mean K (k_μ) as a function of which the deviations from the year (k_t) are derived. The present invention does not model survival rates based on actual survey periods (typically between 2 and 5 years), but rather utilizes different and partial overlap of different beginning and ending survey years equally to generate a potential variable (k_lat) of survival probability K on a logical scale per plot and per year. Thus, the actual binary survival observations of each species per investigation t are used to generate potential probabilities of survival probability contained annually during successive investigation.
The survival Logistic function modeling, i.e. modeling by one of the sigmoid or inverted sigmoid Logistic functions, depends on the size of the respective individual, i.e. whether it is smaller or larger than a threshold value, here set to the average size of the species. The present invention extracts the median value (i.e., the probability of survival of a plant over a given year) of the k_lat posterior distribution for each potential observation. The distribution of median k_lat values for all 81 species and 24 plots was summarized and used as the primary response variable for investigation of time-to-survival patterns in the subsequent model. The present invention focuses on k_lat variation across groupings (species, plot and year) as it best summarizes the mortality of plants during their growth cycle. Although other parameters have an impact on early and late mortality, there is no concern about differential mortality in small individuals given the availability of data. Therefore, the vegetation dead risk probability (hereinafter referred to as death risk), i.e., annual death risk is defined as 1 minus survival probability, and the coefficient sign in the model is reversed (e.g., the negative effect on survival becomes the positive effect on death risk). Note that since the present invention focuses on linking mortality to ontogenesis by size, deviations due to screening interval length deviations, which result from events that the longer interval misses individuals as long as the screening diameter and dies before being recorded, cannot be taken into account. Since these deviations increase mortality estimates for shorter field investigation interval lengths and have a more pronounced impact on smaller size vegetation, the present invention example may actually underestimate mortality after an increase in early check intervals (the 1980 s).
Therefore, the static basic model of the estimation and prediction model of the dead area of grassland vegetation in the permafrost region of Qinghai-Tibet plateau is shown as the equation (2):
wherein K is μ Normal … equation (3)
In the above model, the number of the modules,represents all years, K, in a year chain cycle processing time interval lat Represents the potential survival rate per year under ideal conditions, < ->An inverse Logistic function representing annual survival. Equation (3) represents an a priori function showing average survival.
The model will K lat The overall change between different species, plots and years is broken down into species levels (σ j ) Grade of land block (sigma) K ) And year level (sigma) T ) Deviation from the overall average value (alpha) 0 The method comprises the steps of carrying out a first treatment on the surface of the I.e. average K in the whole dataset lat ). Thus, the model is an unconditional Bayesian multi-level or hierarchical model, in that it contains only varying intercepts (also called "random" intercepts) for modeling K lat (i.e., an overall intercept or a large average). To facilitate the specification of a reasonable priori, K lat Normalized to zero mean and Standard Deviation (SD)In units of.
S3, estimating dead areas of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau and assigning performance parameters of a time change inflection point (change point) of a prediction model;
in actual life, the change of the vegetation survival rate or the vegetation death rate of the grasslands with time is not single rise or fall, but repeatedly fluctuates, and particularly the change of the vegetation survival rate or the vegetation death rate of the grasslands from rise to fall or from fall to rise can be caused by extreme warmth or drought in some years. Therefore, the invention carries out time change inflection point (change point) performance assignment on the estimation and prediction model (equation (2)) of the dead area of the grassland vegetation in the multi-year frozen soil region of the Qinghai-Tibet plateau.
First, an unconditional, multi-level Bayesian model K_lat was generated for all 81 species to visually evaluate the temporal trend of average annual survival rate, and varying intercept (also referred to as "random" intercept) was used to model risk of death (i.e., K_lat; i.e., overall intercept or large mean) and deviation of species, plot and grade from large mean. Because the increasing trend of the vegetation dead risk of the grassland in the 80 th year of the 20 th century is obvious, the invention uses a model to test the year of the change point.
Second, because the present invention uses a change point analysis to statistically verify whether the average value of the time series of potential logical survival probabilities (K_lat) has changed, if so, the year in which such a change is most likely to occur is determined. The point of time-trend in mortality risk was found in 1984 from slight negative to apparent positive in all plots and species. All further analyses focused on the structure and underlying causes of increased risk of vegetation blight in grasslands during the 35 years 1984-2019. Due to the time lag between the cause of death and the event of death, vegetation death is often due to a variety of mechanisms, as well as delays between the event of death and its detection, and it is not possible to directly establish a causal relationship between vegetation death and the average climate conditions for a particular weather event or census interval. Thus, the present invention uses different data sets and methods to best relate patterns identified in long-term data to known vegetation-blight potential drivers.
Therefore, the invention estimates and predicts the parameter K in the basic static model for the dead area of grassland vegetation in the permafrost region of the Qinghai-Tibet plateau P ，K_T，∈σ P ，σ T ，σ ∈ And p1 (see specifically equations (4) to (8)) so that the estimation and prediction model (equation (2)) of dead areas of grassland vegetation in the multi-year frozen soil region of Qinghai-Tibet plateau can capture inflection points or change points in real production, namely, the change points of death risk time trend from slight negative to obvious positive are found in all plots and species in 1984.
The average survival rate after adjustment according to the sample side information is expressed as a priori:
K P ～Normal(0，σ P ) … equation (4)
The average survival rate adjusted according to time information (annual scale) a priori is expressed as:
K_T～Normal(0，σ T ) … equation (5)
The residual after adjustment according to the existing information is expressed as:
∈～Normal(0，σ ∈ ) … equation (6)
The a priori representation of the standard error coefficient is:
σ P ，σ T ，σ ∈ gamma (3, 4) … equation (7)
The a priori representation of the p1 survival parameters is:
P1-Normal (min (dbh), 1) … equation (8)
S4, objectively evaluating experience of a vegetation dead area estimation and prediction model in a permafrost region of the Qinghai-Tibet plateau;
because the invention solves the actual problem in actual production activities in permafrost regions of Qinghai-Tibet plateau, the reliability or objectivity of the model must be assigned. Meanwhile, to explore the effect of differences in mortality risk and long-term average climate and the location of species on their climatic habitat, the present invention integrates the time trends of Tmax, VPD and MCWD between 1975 and 2023. For this purpose, a bayesian generalized additive model is used, the month values of the three climate variables are modeled with B-splines and up to four basis functions, respectively, to limit the volatility of the relationship and to add varying intercepts for the different sample parties. The results show a trend that shows a significant increase in Tmax and VPD over time in all samples, but no increase in MCWD.
The vegetation survival rate model is based on a Bayesian hierarchical model, so that the vegetation survival rate model is divided into three layers, namely a sample side level, a species and a species interior, and the dead time change of grassland vegetation is specifically defined as follows:
1) Sample formula level: the increase in vegetation dead risk in the sample after 1984, tests showed that the proportion of samples with increased risk of death, and whether the average risk of death and the risk of death over time were dependent on the local average climate. For this purpose, the invention uses a party mean climate factor, defined as a 35 year local mean of climate variables (1984-2019, matching the main period of interest). The effect of the VPD 30 year average in the drier year and the maximum temperature (Tmax) in the warmest year are also particularly considered, as these two variables increase significantly in all 24 samples over the past 49 years and are important driving factors for vegetation growth in these samples. In addition, the effect of MCWD, an index that is a proxy for the annual accumulated moisture pressure in arid seasons, was also studied, estimated as an accumulated deficit between precipitation and transpiration. These variables may all affect survival by compromising moisture pressure or biochemical reactions associated with heat. The transpiration pressure and temperature data were collected from ancillimate v.2.0 in 1971 to 2019, and MCWD was calculated using precipitation data from ancillimate v.2.0 and transpiration data from terralclate.
2) Inter-species variation: inter-species variation uses a bayesian multilayer model similar to intra-sample vegetation mortality estimation, but does not include climate variables, in order to focus on species-specific mortality risk variation trajectories and test the proportion of species that increase mortality risk over time. The present invention then tested whether the species had a lower average survival rate in plots near the dry edge of their geographic range (upper limit of their total VPD habitat) as a significant increase in temperature and VPD was likely to drive an increase in the risk of vegetation dead. Likewise, MCWD is used to approximate the rainfall seasonal habitat of a species. This approach takes into account the general ecological physiological constraints and evolutionary history of each species, as by captured conditions of the geographical extent of their occurrence, for example, a difference in average temperature of two plots of 2 ℃ may be associated with a higher minimum dead rate for one species, since for that species, warmer plots are close to the edge of their temperature habitat, while for the other species, the minimum mortality remains unchanged if both plots are close to their habitat center. This biophysical habitat approach allows for the study of potential mechanisms that may be the basis of temporal trends, while also providing potential ecological physiological mechanisms that are the basis of temporal and spatial patterns.
To calculate the approximately univariate climatic habitat for each species and to determine the relative positions of the investigated parties in these habitats, all known occurrence records for 78 species were first extracted from the global biodiversity information facility online database. The goal is to strike a balance between considering only the simplicity of presence analysis and avoiding the prejudice that may underestimate the width of the habitat. Therefore, the invention first filters out multiple occurrences of the same species, leaving only one sample within each 1 km radius to reduce the risk of over-sampling. The filtered locations of occurrence of the 78 species were matched to a 1/24 ° 30 year climatic chart (average in 1981-2010), including maximum year VPD, maximum day air temperature (Tmax) and MCWD, from terrraclimate. The univariate climatic habitats of the species are defined separately for Tmax, VPD and MCWD, respectively, as a distribution of 30 years averages extracted from pixels corresponding to the biogeography of each species, including the plots of the present invention.
The local 30-year average climate data for 24 sample parties is then expressed in terms of quantiles based on the specific climate habitat of each species. This resulted in Tmaxniche, VPDniche and MCWDniche, i.e. the location of the corresponding variable of the species in its total climatic habitat. The resulting score is a species-specific expression, indicating how close the 30-year-average climate variable in the plot of the invention is to the upper limit of the total climate habitat of the species. The robustness of the quantile method against partial imperfect sampling in the biological geographical area was also tested by re-running the M5 model using a coarse index whose distance from the habitat maximum is expected to be less sensitive to non-uniform sampling efforts in the total distribution area of the species.
3) Species layer: in order to explore the underlying physiological ecological mechanisms that affect species level survival changes after 1984, the present invention only considered 40 species with functional trait data, not all 81 species. Functional trait is a measurement of 81 plants from seven of 24 samples during 2015, 7 months to 9 months. For each party, the goal of the selection of species was to sample those on site that constituted 80% of the last census. The trait of each species was measured on three individuals, including at 400. Mu. Mol -1 Reference CO of (2) 2 Concentration and 1,500. Mu. Mol photons m -2 s -1 Leaf photosynthesis and stomatal conductance (Asat and gsat, respectively) under illumination of (a) at the same CO 2 Dark breathing (Rd) at concentration, CO 2 Saturated photosynthesis and stomatal conductance (Amax and gmax, respectively), at 1,200. Mu. Mol -1 CO 2 And (5) measuring.
The A-ci curve fitted using the photosynthesis model in the "plantecophys" R package yields estimates of the maximum carboxylation rate (Vcmax) of the leaf and the maximum light driving electron flow (JMax) normalized to 25 ℃. A one-point method was used to obtain a reaction product of 400. Mu. Mol -1 CO 2 The net photosynthesis measured below estimates Vcmax for each individual, and from that measured at 1,200. Mu. Mol -1 CO 2 The net photosynthesis measured below estimates Jmax. In addition, leaf area, leaf mass ratio, leaf thickness, stem density, leaf nutrient concentration, and stable carbon isotope ratio (delta) were measured 13 C) A. The invention relates to a method for producing a fibre-reinforced plastic composite All traits were averaged at the species level.
Therefore, in order to meet the above definition, the parameters p2, r1 and r2 of the estimation and prediction model (equation (2)) of the dead-land vegetation in the permafrost region of the Qinghai-Tibet plateau are assigned, see equations (9) to (11), so that the estimation and prediction model of the dead-land vegetation in the permafrost region of the Qinghai-Tibet plateau can accurately estimate the dead-land time variation trend of the vegetation in the grassland vegetation, whether at the sample side level, between species and within species.
The a priori representation of the p2 survival parameters is:
P2-Normal (max (abh), 1 … equation (9)
The a priori representation of the r1 survival parameters is:
r 1-Uniform (0, 2) … equation (10)
The a priori representation of the r2 survival parameters is:
r 2-form (-2, 0) … equation (11)
S5, estimating dead area of grassland vegetation in permafrost areas of the Qinghai-Tibet plateau and testing and verifying stability of a prediction model;
in order to test whether a grassland vegetation dead area estimation and prediction model of a Qinghai-Tibet plateau for many years works stably, namely, the degree of influence (sensitivity) of the robustness of the time trend of the death risk increase after the disaster disturbance related to the freeze injury is eliminated or the model accuracy rate by the extreme event, the sensitivity analysis is carried out:
first, two data subsets of the initial raw survival dataset are created by filtering out any census intervals affected by freeze injury, or any samples that have undergone at least one freeze injury disturbance since 1975, even samples that have been slightly freeze injury. The resulting dataset contained 23 and 5 samples, covering 69 and 15 species, respectively, exceeding 400 observations.
These two data sets are then separately input into equation (1), generating a potential annual survival probability (k_lat) of vegetation for each species separately, and a model is run to test the sample and species-level risk of death time trends.
The results of all sensitivity analysis show that even after the main influence of disturbance related to freeze injury is eliminated, the death risk trend between the sample side and the species is still obvious, the stability of the model reaches more than 83 percent, the sensitivity is less than 12 percent, and the requirements of actual production, life and work are met.
S6, estimating dead area of grassland vegetation in permafrost areas of the Qinghai-Tibet plateau and updating parameters of a prediction model:
the parametric probability posterior distribution of the grassland vegetation dead area estimation and prediction model in the perennial area of the Qinghai-Tibet plateau was fitted separately for each species in Stan using No-U-Turn-Sampler (NUTS; 5,000 iterations on three chains) with the 'rstan' R package. Bayesian updates to the grassland vegetation dead area estimation and prediction model parameter probability distribution in the perennial soil region of the Qinghai-Tibet plateau were fitted using NUTS in Stan by running 6,000 iterations on four chains, including 1,000 "warm-up" steps, using the R-package "brms". The chains were checked for convergence by the Rhat value, ensuring that they were all within 0.01 around 1, and the mix of all chains was visually assessed. The posterior distributions of coefficients are summarized by their median and 95% of the highest posterior density interval HPDI (i.e., the narrowest posterior interval containing 95% probability mass, corresponding to the coefficient value most consistent with the data). When the coefficients of the model covariates 95% HPDI did not contain zero, they were considered important, indicating a sufficiently strong confidence level to report the effect as either positive or negative.
Claims (1)
1. A method for estimating and predicting dead areas of grasslands in permafrost regions of Qinghai-Tibet plateau comprises the following steps:
s1, data arrangement is carried out on environmental data of permafrost areas of Qinghai-Tibet plateau and vegetation species of grasslands to form a data set;
s2, establishing a static basic model of a year-year dead area estimation and prediction model of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau by using a Logistic function, wherein the model is as follows:
K μ ～normal(2，1)
wherein:the method comprises the steps of (1) circularly processing a year chain, namely, finishing and forming all years in a time interval of environmental data of a permafrost region of a Qinghai-Tibet plateau and a grassland vegetation species data set in the step of S1; k (K) lat Represents the potential survival rate per year under ideal conditions; />An inverse Logistic function representing annual survival rate; k (K) μ Normal (2, 1) represents an a priori function showing average survival; k (K) μ The intrinsic survival rate of a specific species, namely the survival rate which is free from external environmental factors and time; />Represents the p-like average survival rate of the ith species; />Represents the average survival rate of the ith species in the x year; epsilon i Representing an ith species viability fit residual; θ represents the grassland vegetation survival probability; DBH (DBH) i Represents the average height of adult plants of the i-th species; threshold represents a fitting threshold demarcation point; k is the survival probability of most of the individual's adulthood; r1 is the rate at which the survival rate of the S-shaped portion increases to reach an arbitrary size point at which the survival probability becomes K; p1 is the size corresponding to the inflection point of the functional form of the S-shaped portion; r2 is the rate of decline in annual survival rate of species of the inverted S-shaped portion approaching its maximum size; p2 is the inflection point or threshold height or size corresponding to 50% of the annual survival probability of the inverted S-shaped portion;
s3, estimating dead area of grassland vegetation in a permafrost region of the Qinghai-Tibet plateau and assigning a time change inflection point performance parameter of a prediction model:
the average survival rate after adjustment according to the sample side information is expressed as a priori:
K P ～Normal(0，σ P )；σ P is p-like standard difference;
the average survival rate after adjustment according to the time information is a priori expressed as:
K_T～Normal(0，σ T )；σ T standard deviation of time T;
the residual after adjustment according to the existing information is expressed as:
∈～Normal(0，σ ∈ )；σ ∈ standard deviation for residual e;
the a priori representation of the standard error coefficient is:
σ P ，σ T ，σ ∈ ～Gamma(3，4)；
the a priori representation of the p1 survival parameters is:
p1～Normal(min(dbh)，1)；
s4, objectively evaluating experience of a vegetation dead area estimation and prediction model in a permafrost region of the Qinghai-Tibet plateau;
the a priori representation of the p2 survival parameters is:
p2～Normal(max(dbh)，1)；
the a priori representation of the r1 survival parameters is:
r1～Uniform(0，2)；
the a priori representation of the r2 survival parameters is:
r2～Uniform(-2，0)；
s5, estimating the dead area of grassland vegetation in the permafrost region of the Qinghai-Tibet plateau and updating the parameters of the prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410088851.XA CN117787503B (en) | 2024-01-22 | 2024-01-22 | Estimation and prediction method for dead area of grassland in permafrost region of Qinghai-Tibet plateau |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410088851.XA CN117787503B (en) | 2024-01-22 | 2024-01-22 | Estimation and prediction method for dead area of grassland in permafrost region of Qinghai-Tibet plateau |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117787503A true CN117787503A (en) | 2024-03-29 |
| CN117787503B CN117787503B (en) | 2024-05-28 |
Family
ID=90400126
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410088851.XA Active CN117787503B (en) | 2024-01-22 | 2024-01-22 | Estimation and prediction method for dead area of grassland in permafrost region of Qinghai-Tibet plateau |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117787503B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106126484A (en) * | 2016-07-06 | 2016-11-16 | 中交第公路勘察设计研究院有限公司 | The multi-factor comprehensive ever-frozen ground ground temperature zoning methods of multiple linear regression analysis |
| CN107330279A (en) * | 2017-07-05 | 2017-11-07 | 贵州省草业研究所 | A kind of high mountain permafrost area vegetation pattern Forecasting Methodology |
| WO2022135265A1 (en) * | 2021-01-14 | 2022-06-30 | 中国长江三峡集团有限公司 | Failure warning and analysis method for reservoir dispatching rules under effects of climate change |
-
2024
- 2024-01-22 CN CN202410088851.XA patent/CN117787503B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106126484A (en) * | 2016-07-06 | 2016-11-16 | 中交第公路勘察设计研究院有限公司 | The multi-factor comprehensive ever-frozen ground ground temperature zoning methods of multiple linear regression analysis |
| CN107330279A (en) * | 2017-07-05 | 2017-11-07 | 贵州省草业研究所 | A kind of high mountain permafrost area vegetation pattern Forecasting Methodology |
| WO2022135265A1 (en) * | 2021-01-14 | 2022-06-30 | 中国长江三峡集团有限公司 | Failure warning and analysis method for reservoir dispatching rules under effects of climate change |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117787503B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chipanshi et al. | Evaluation of the Integrated Canadian Crop Yield Forecaster (ICCYF) model for in-season prediction of crop yield across the Canadian agricultural landscape | |
| Clark et al. | Stages and spatial scales of recruitment limitation in southern Appalachian forests | |
| Uriarte et al. | Multidimensional trade‐offs in species responses to disturbance: implications for diversity in a subtropical forest | |
| Guse et al. | On characterizing the temporal dominance patterns of model parameters and processes | |
| Jorgensen et al. | Assessing landscape constraints on species abundance: does the neighborhood limit species response to local habitat conservation programs? | |
| Saracco et al. | Assessing landbird monitoring programs and demographic causes of population trends | |
| Cipriotti et al. | Understanding the long‐term spatial dynamics of a semiarid grass‐shrub steppe through inverse parameterization for simulation models | |
| Roth et al. | Functional ecology and imperfect detection of species | |
| Perrone et al. | The relationship between spectral and plant diversity: Disentangling the influence of metrics and habitat types at the landscape scale | |
| Forcey et al. | Spatially explicit modeling of blackbird abundance in the Prairie Pothole Region | |
| He et al. | Data‐driven estimates of global litter production imply slower vegetation carbon turnover | |
| Rubinstein et al. | An eco-spatial index for evaluating stabilization state of sand dunes | |
| Jenouvrier et al. | Detecting climate signals in populations across life histories | |
| CN117787503B (en) | Estimation and prediction method for dead area of grassland in permafrost region of Qinghai-Tibet plateau | |
| Lin et al. | Identifying the spatial mixture distribution of bird diversity across urban and suburban areas in the metropolis: A case study in Taipei Basin of Taiwan | |
| Xu et al. | [Retracted] Digital Technology Empowers Grain Supply Chain Optimization Simulation | |
| Riva et al. | Of detectability and camouflage: evaluating Pollard Walk rules using a common, cryptic butterfly | |
| Signer et al. | Detecting the influence of environmental covariates on animal movement: a semivariance approach | |
| Hemstrom | Late-successional and old-growth forest effectiveness monitoring plan for the Northwest Forest Plan | |
| Gao et al. | Ecological risk assessment of grassland vegetation change based on Bayesian model in Xilin Gol League, China | |
| Chaudhary et al. | Deep Learning Approach for Forecasting Apple Yield Using Soil Parameters | |
| Gamelon et al. | Detecting climate signals cascading through levels of biological organization | |
| Bayne et al. | Adapting monitoring to more effectively assess the impacts of climate change on Alberta's biodiversity | |
| Escoriza et al. | Using hierarchical spatial models to assess the occurrence of an island endemism: the case of Salamandra corsica | |
| Dai et al. | Large-scale spatiotemporal deep learning predicting urban residential indoor PM2. 5 concentration |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |