=Paper=
{{Paper
|id=Vol-3006/54_short_paper
|storemode=property
|title=Long-term forecast of heavy metals content in wheat grain under changing climate conditions
|pdfUrl=https://ceur-ws.org/Vol-3006/54_short_paper.pdf
|volume=Vol-3006
|authors=Yuri B. Kirsta,Alexander V. Puzanov,Tamara A. Rozhdestvenskaya,Maria P. Peleneva
}}
==Long-term forecast of heavy metals content in wheat grain under changing climate conditions==
https://ceur-ws.org/Vol-3006/54_short_paper.pdf
Long-term forecast of heavy metals content in wheat
grain under changing climate conditions
Yuri B. Kirsta1 , Alexander V. Puzanov1 , Tamara A. Rozhdestvenskaya1 and
Maria P. Peleneva1
1
Institute for Water and Environmental Problems of Siberian Branch of the Russian Academy of Sciences, Barnaul, Russia
Abstract
Using the system approach, we have developed a simulation model for the long-term forecast of the
content of toxic chemical elements in grain crop yield. The study was carried out by the example of
wheat cultivated in Altai Krai — one of the main grain-producing regions of Russia. Wheat crops were
sampled in 10 municipal districts of Altai Krai, which characterize seven different edaphic-climatic
zones. The average long-term values of mean monthly air temperature and monthly precipitation for
each sampling area were identified using GIS and data of the Interactive Agricultural Ecological Atlas
of Russia and Neighboring Countries. A total of 19 chemical elements were considered, i.e. Pb, As,
Cd, Hg, Na, Mn, Zn, Cu, Fe, Co, etc. It is shown that content of Pb, Na, Mn and Cu in wheat depend
on climatic characteristics of the cultivation area. Regression dependences of element content on the
average long-term air temperature and precipitation were established. Based on normalization and
spatial generalization of air temperature and precipitation providing the uniform dynamics of their
relative monthly values (in percent) throughout the study area, a forecast of their changes was made for
2030. A procedure for grain sampling, GIS technologies for processing meteorological and cartographic
data, methods for predicting regional climate changes and establishment of quantitative relationships
of chemical elements content in grain with climatic characteristics – all together make up the integral
predictive simulation model for toxic substance content in grain crop yield. The model was used for
estimation of Pb, Na, Mn, Cu changes in wheat by 2030. The lead (Pb) content in wheat crop delivered to
elevators from certain municipal districts will exceed the maximum allowable concentration for bread-
grain after 2030. Unlike Pb, Na, Mn, Cu, the content of other metals in wheat grain weakly correlate with
long-term changes in air temperature and precipitation; therefore, it can hardly change significantly.
Keywords
Cereals, wheat, heavy metals, forecast, climate change, Altai.
1. Introduction
Our research is devoted to the development of a methodology for simulation modeling and
forecasting the heavy metals content in wheat grain by 2030 with minimum data of observations.
We take as an example the Altai Krai that is one of the largest grain-producing regions of the
Russian Federation. The heavy metals content depends on the climate and the hard-to-consider
local conditions of crop cultivation area, including local soil characteristics, rotation crops,
chemicals used and many other factors. Remaining within the framework of the system approach,
we will use statistical methods of data processing that are traditionally used in forecasting poorly
SDM-2021: All-Russian conference, August 24–27, 2021, Novosibirsk, Russia
" kirsta@iwep.ru (Y. B. Kirsta)
© 2021 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
Workshop
Proceedings
http://ceur-ws.org
ISSN 1613-0073 CEUR Workshop Proceedings (CEUR-WS.org)
463
Yuri B. Kirsta et al. CEUR Workshop Proceedings 463–470
studied complex processes. Wheat cultivation technologies and soil characteristics affecting the
quality of grain remain stable over long periods, in contrast to modern climate changes. It is
the influence of the latter that we will analyze.
Currently, a large number of empirical-statistical and deterministic methods of long-term
meteorological forecasts, as well as their various combinations, are proposed [1, 2, 3, 4]. The
first ones are based on the statistical regularities of atmospheric processes and require the
maximum volume and homogeneity of the studied series of meteorological data. The latter are
founded on the physical laws of atmospheric or atmospheric-oceanic processes and describe
them by complex equations as, for example, in the mesoscale WRF model [5, 6].
2. Basic data
When developing universal predictive models, characterizing long-term changes in chemical
elements content in cereals, it is necessary to move from specific units of substance measurement
to dimensionless characteristics [7, 8]. We normalized the observational data of each element
content to its average for the Altai Krai (for all sampling municipal districts characterizing
seven edaphic-climatic zones) and expressed them as a percentage of this average (Table 1).
To assess the climate situation in the Altai Krai and adjacent territories, we used 11 reference
weather stations with observations 1951–2020 [10, 11]. According to the developed method
Table 1
Chemical elements content in wheat grain in 2018–2019, normalized to the average values over the
territory, %.
Chemical
Allowable
element
Average1 ,
Sampling district number limit2 ,
mg/kg
1 2 3 4 5 6 7 8 9 10 mg/kg
Pb 34.5 37.53 56.3 133.17 133.17 169.49 127.12 133.17 96.85 78.69 0.1652 0.5
Cd 45.05 103.6 108.11 45.05 148.65 144.14 45.05 108.11 117.12 135.14 0.0222 0.1
Ca 99.52 102.21 99.52 107.58 10.22 115.65 110.27 102.21 134.48 118.34 3718 −
K 109 85.31 87.68 90.05 99.53 113.74 97.16 130.33 99.53 87.68 4.22 −
Mg 107.96 106.13 96.07 100.18 110.7 150.96 90.58 135.41 13.08 88.75 1093 −
Na 76.82 69.14 57.62 69.14 172.86 179.26 96.03 140.85 69.14 69.14 78.1 −
P 98.67 90.67 88 88 88 114.67 117.33 117.33 106.67 90.67 3750 −
Mn 91.06 106.24 80.94 97.81 106.24 111.3 97.81 119.73 96.12 92.75 59.3 −
Zn 125.62 152.71 76.35 113.3 105.91 105.91 91.13 76.35 88.67 64.04 40.6 −
Cu 106.64 117.13 78.67 111.89 110.14 106.64 104.9 115.38 82.17 66.43 5.72 −
Fe 116.07 133.93 108.63 84.82 111.61 114.58 92.26 78.87 83.33 75.89 67.2 −
Co 9.171 9.171 9.171 148.62 89.91 104.59 177.98 201.83 161.47 88.07 0.0545 −
Al 56.95 125.28 63.78 116.17 100.23 125.28 184.51 50.11 111.62 66.06 43.9 −
Ba 150.63 87.87 125.52 83.68 79.5 125.52 58.58 125.52 71.13 92.05 2.39 −
Mo 40.12 19.55 56.58 95.68 22.63 59.67 339.51 87.45 226.34 52.47 0.972 −
Ni 35.61 83.09 68.25 127.6 130.56 94.96 97.92 225.52 80.12 56.38 0.337 −
1
We normalized the value of 0.005 mg/kg as half of the element detection limit of 0.01 mg/kg.
2
Allowable level in bread-grain, mg/kg [9].
464
Yuri B. Kirsta et al. CEUR Workshop Proceedings 463–470
of normalization and spatial generalization of meteorological data [7, 12, 13, 14], the observed
values of air temperature and precipitation for each month in each year were recalculated as
a percentage relative to their “in situ” average long-term values for January and July. Such
normalized values allow a uniform description of long-term changes in temperature and precip-
itation over plain and mountainous areas. To switch back to the generally accepted units of
factor measurements (∘ C, mm), it is enough to know their average long-term January and July
values in the characterized location.
3. Forecast of changes in air temperature and precipitation by
2030
The system analysis of climate dynamics for the grain-producing zone of Russia showed that this
dynamics obeys an age-long climatic cycle consisting of three 33-year phases [11, 15, 16]. These
phases cover 1918–1950, 1951–1983, 1984–2016, and each of them is characterized by certain
statistical regularities of long-term changes in normalized air temperature and precipitation.
To predict these factors by 2030, we used the meteorological patterns of the third phase as
the closest to the forecast period. We normalized and spatially generalized (averaged over
11 reference weather stations) the monthly values of these factors for 1984–2020. Then, using
the obtained data, we calculated the long-term linear trends of factors for each month of the
year. Extrapolating the found trends for 10 years ahead, we obtained normalized monthly values
of air temperature and precipitation expected by 2030.
The forecast shows that by 2030 the average monthly air temperature in January will decrease
(∼ 20%), in March and April will increase (> 20%), and will remain approximately the same
in the other months. The predicted changes in monthly precipitation are multidirectional by
months of the year. Note that the forecast of normalized climatic characteristics is uniform
throughout the Altai Krai, regardless of the orographic and climatic heterogeneity of its territory.
4. Relationship between heavy metals content in wheat grain
and climatic conditions
The short observation period 2018–2019 for chemical elements content in wheat grain (Table 1)
does not allow us to find its dependence on climatic factors directly. Because of this, we
compared this content with the average long-term values of mean monthly air temperature and
monthly precipitation corresponding to the sampling municipal districts. Each of 10 sampling
districts (Table 1) has its own average long-term values of climatic factors. As a result, it becomes
possible to determine the relationship between the substance content in grain and these factors.
To estimate the average long-term monthly values of air temperature and precipitation for
each sampling district, we used the Interactive Agricultural Ecological Atlas of Russia and
Neighboring Countries [17]. We calculated these values by means of the GIS platform ArcGIS
Desktop 10.0. Then we normalized and spatially generalized them over the territory of the Altai
Krai. The data obtained made it possible to link the substance content in the grain with climatic
factors.
465
Yuri B. Kirsta et al. CEUR Workshop Proceedings 463–470
To describe most adequately the substance content in wheat grain as a function of tempera-
ture and precipitation, we first determined the influence of their monthly values separately. For
this purpose, the correlation between the normalized content of each chemical element and the
average long-term normalized values of mean monthly air temperature and monthly precipita-
tion was found. Then we identified the months that gave the highest correlation coefficients
for each meteorological factor. Precipitation showed a more significant impact if compared
to air temperature. It turned out that the chemical elements content in grain significantly
depends on the meteorological factors of winter period. Combining the months found for each
meteorological factor, we calculated the final correlation coefficients between the chemical
elements content and the factors (Table 2). In Table 2, the largest correlation coefficients are
shown in bold.
Choosing the chemical elements Pb, Na, Mn, Cu, which are most dependent on meteorological
factors (Table 2), we determined the relationship of their content in wheat grain with the
average long-term normalized values of air temperature and precipitation, that is, with climatic
conditions in the Altai Krai. For each found linear relationship, we also calculated the reliability
of approximation 𝑅2 (coefficient of determination). The linear dependences Pb ([𝑃1 + 𝑃2 +
· · · + 𝑃12 ]/12), Na ([𝑃1 + 𝑃2 + · · · + 𝑃12 ]/12), Mn ([𝑃7 + 𝑃8 ]/2), Cu ([𝑃5 + 𝑃6 + 𝑃7 ]/3),
Pb ([𝑇5 + 𝑇6 + 𝑇7 + 𝑇8 ]/4), Na ([𝑇5 + 𝑇6 + 𝑇7 + 𝑇8 ]/4), Cu ([𝑇12 + 𝑇1 + 𝑇2 + 𝑇3 ]/4) showed
the sufficiently high values of 𝑅2 : 0.59, 0.66, 0.52, 0.51, 0.68, 0.61, 0.63, correspondingly.
Different soil characteristics and other agricultural factors affect the scatter of data relative
to the regression lines and reduce the coefficient of determination 𝑅2 . At the same time, the
Table 2
Correlation coefficients between the normalized chemical elements content in wheat grain (Table 1) and
the average long-term normalized monthly values of air temperature (𝑇𝑖 , 𝑖 = 1, . . . , 12) and precipitation
(𝑃𝑖 , 𝑖 = 1, . . . , 12).
Chemical 𝑇5 + 𝑇6 + 𝑇7 + 𝑇8 𝑇12 + 𝑇1 + 𝑇2 + 𝑇3 𝑃1 + 𝑃2 + · · · + 𝑃12 𝑃5 + 𝑃6 + 𝑃7 𝑃7 + 𝑃8
element 4 4 12 3 2
Pb 0.8236 −0.3222 −0.7682 −0.6535 −0.7442
Cd 0.1776 0.2511 −0.1450 −0.0677 −0.1636
Ca −0.2649 0.5313 0.4699 0.4902 0.3927
K 0.4964 −0.2482 −0.4559 −0.4663 −0.5913
Mg 0.6432 −0.6754 −0.6700 −0.6730 −0.6142
Na 0.7812 −0.4536 −0.8099 −0.7361 −0.8151
P 0.4518 0.0611 −0.2239 −0.1941 −0.3643
Mn 0.5612 −0.3372 −0.5647 −0.5679 −0.7191
Zn −0.0930 −0.4645 −0.1322 −0.2132 −0.1756
Cu 0.4808 −0.7967 −0.6196 −0.7118 −0.7127
Fe 0.0464 −0.4316 −0.1726 −0.2483 −0.1673
Co 0.4089 0.0613 −0.2759 −0.2188 −0.3647
Al 0.2769 −0.1029 −0.1911 −0.1380 −0.1769
Ba 0.0848 −0.2507 −0.1316 −0.1829 −0.1309
Mo 0.0560 0.2901 0.1603 0.1914 0.1086
Ni 0.6028 −0.4197 −0.5612 −0.6224 −0.6971
466
Yuri B. Kirsta et al. CEUR Workshop Proceedings 463–470
nature of soils in agroecosystems cannot change significantly over a decade. Considering this
and “in situ” normalization of meteorological factors, we can use the found dependences to
predict changes in the chemical elements content in wheat grain by 2030. “In situ” soil features
therewith have little effect on the reliability of such forecast.
5. Forecast of heavy metals content in wheat grain for 2030
The following simulation linear equation is used to predict the chemical elements content in
grain crop yield:
𝐹 𝐶 = 𝐶𝐶 + 𝑘 × (𝑀 𝑉 30 − 𝑀 𝑉 ),
where 𝐹 𝐶 is the predicted content of chemical element (Pb, Na, Mn, Cu) in wheat grain by 2030;
𝐶𝐶 is the current value of the content (2018–2019); 𝑘 is the slope coefficient of linear regression
for the element content depending on climatic conditions; 𝑀 𝑉 is “in situ” average long-term
monthly values of precipitation or air temperature: [𝑃1 + 𝑃2 + · · · + 𝑃12 ]/12, [𝑃5 + 𝑃6 + 𝑃7 ]/3,
[𝑃7 + 𝑃8 ]/2, [𝑇5 + 𝑇6 + 𝑇7 + 𝑇8 ]/4, [𝑇12 + 𝑇1 + 𝑇2 + 𝑇3 ]/4; 𝑀 𝑉 30 — the analogous values of
precipitation or air temperature that are predicted for 2030. The forecast equation characterizes
the increase or decrease in the chemical elements content in grain, depending on the change
(𝑀 𝑉 30 − 𝑀 𝑉 ) of climatic characteristic.
The 2030 forecast on the Pb, Na, Mn, Cu content in wheat grain showed that the influence of
air temperature during growth season is noticeably stronger than of precipitation. The influence
of winter air temperature [𝑇12 + 𝑇1 + 𝑇2 + 𝑇3 ]/4, which determines the degree of freezing of
soils, the intensity of their thawing in spring and, hence, the development of cereal seedlings, is
also significant (Figure 1).
The forecast (Figure 1) allows us to estimate critical changes in the chemical elements
content in wheat crop yields by 2030. Consider the strong dependence of Pb on air temperature
[𝑇5 + 𝑇6 + 𝑇7 + 𝑇8 ]/4 during the growth season May-August. This dependence leads to a
significant increase in the Pb content up to 200–250% in 4th , 5th , 6th , 7th and 8th districts of the
Altai Krai. These are Romanovsky, Rodinsky, Klyuchevskoy, Aleisky (steppe) and Yegoryevsky
districts [10]. Converting percentages into units of substance measurement mg/kg (see Table 1),
we find that, for example, in Klyuchevskoy district (No. 6) the Pb content in grain by 2030 will
reach the following value:
(Average Pb content) × 250/100 = 0.1652 × 2.5 = 0.41 mg/kg.
The resulting value is critically close to the allowable level of Pb in food wheat, which is
0.5 mg/kg (Table 1). After 2030, the Pb content in some grain shipments delivered to grain
elevators will exceed this level due to the natural statistical variation of this characteristic. The
content of other chemical elements in wheat grain (Table 1) will change insignificantly and not
prevent its use for food purposes.
The presented forecast model of the chemical elements content in grain yield is based on the
system analysis and mathematical simulation modeling of climatic and agroecosystem processes.
It includes predicted trends in air temperature and precipitation, linear regressions for chemical
elements, the content of which in cereals is sensitive to changes in climatic conditions. The
467
Yuri B. Kirsta et al. CEUR Workshop Proceedings 463–470
Figure 1: The content of lead Pb, sodium Na, and copper Cu in wheat crop yields in 10 sampling
municipal districts of the Altai Krai as a percentage of the average values for the territory (Table 1).
•
( ) — measured in 2018–2019, (■) — predicted for 2030, 𝑇𝑖 — average long-term temperature in the
month 𝑖, 𝑖 = 1, . . . , 12.
method for constructing climatic trends is supported by successful long-term forecasts of air
temperature, precipitation and wheat crop yields in the USA, Russia, the Siberian Federal District
and the Altai Krai [10, 12, 15, 18]. The adequacy of linear regression equations for chemical
elements content is confirmed by significant values of their determination coefficient 𝑅2 . Thus,
the developed model provides an adequate forecast of the chemical elements content in wheat
grain for 2030. It is well established that the content of lead (Pb) in some wheat shipments
delivered to grain elevators after 2030 will exceed the maximum allowable concentration for
bread-grain.
468
Yuri B. Kirsta et al. CEUR Workshop Proceedings 463–470
Acknowledgments
The work was carried out within the framework of the Research Program of the Institute
for Water and Environmental Problems SB RAS (Project 1021032424138-9) with the financial
support of the Russian Foundation for Basic Research (grant No. 18-45-220019-r_a).
References
[1] Parton K.A., Crean J. Review of the literature on valuing seasonal climate forecasts in
Australian agriculture. Report for the Project Improved Use of Seasonal Forecasting to
Increase Farmer Profitability. Australia, Orange: NSW DPI, 2016.
[2] Siegert S., Bellprat O., Menegoz M. et al. Detecting improvements in forecast correlation
skill: Statistical testing and power analysis // Monthly Weather Review. 2017. Vol. 145(2).
P. 437-450.
[3] NCEP Generated Products. In: National Centers for Environmental prediction. 2020. Avail-
able at: https://www.nco.ncep.noaa.gov/pmb/products (accessed March 10, 2020).
[4] Kirsta Y.B., Lovtskaya O.V. Annual range of temperature and precipitation forecast for
Altai-Sayan mountain country // Environmental Dynamics and Global Climate Change.
2020. Vol. 11. No. 1. DOI:10.17816/edgcc34020.
[5] Skamarock W.C., Klemp J.B., Dudhia J. et al. Description of the Advanced Research WRF.
Version 3. NCAR Technical Note NCAR/TN-475+STR. 2008. 520 p. DOI:10.5065/D68S4MVH.
[6] Ignatov R.Yu., Zaichenko M.Yu. et al. Comparison of regional atmospheric model forecasts
under different initial and boundary conditions // Russian Meteorology and Hydrology.
2019. Vol. 44: P. 378–383.
[7] Kirsta Y.B., Lovtskaya O.V. Spatial year-ahead forecast of air temperature and precipitation
in large mountain areas // SN Appl. Sci. 2020. Vol. 2. No. 1044. DOI:10.1007/s42452-020-
2861-6.
[8] Kirsta Y.B., Puzanov A.V. System-analytical modelling: 1. Development of regional models
for mountain river runoff // Eurasian Journal of Mathematical and Computer Applications.
2020. Vol. 8. Is. 2. P. 69–85. DOI:10.32523/2306-6172-2020-8-2-69-85.
[9] Technical Regulation of the Customs Union “On Safety of Grain” (TR TS 015/2011) adopted
by the CCU Decision No. 874 on 9 December 2011, and came into force on 1 July 2013.
[10] Kirsta Y.B., Kurepina N.Y., Lovtskaya O.V. The forecast of climate and agroclimatic potential
in Altai Krai up to 2020 // Bulletin of Altai State Agricultural University. 2013. No. 1.
P. 27–32.
[11] Kirsta Y.B., Lovtskaya O.V., Puzanov A.V. The forecast of climate changes in Altai-Sayan
mountain country till 2030 // CEUR Workshop Proceedings. 2019. Vol. 2534 P. 114–117.
http://ceur-ws.org/Vol-2534/19_short_paper.pdf
[12] Kirsta Yu.B. System-analytical modelling — Part II: Wheat biotime run and yield for-
mation. Agroclimatic potential, Le Chatelier principle, changes in agroclimatic poten-
tial and climate in Russia and the U.S. // Ecol. Modelling. 2006. Vol. 191. P. 331–345.
DOI:10.1016/j.ecolmodel.2005.05.027.
[13] Kirsta Y.B., Lovtskaya O.V. The forecast of climatic changes in grain-producing areas of
469
Yuri B. Kirsta et al. CEUR Workshop Proceedings 463–470
Siberia and Russia // World Sci. Cult. Educ. (Mir Nauki, Kul’tury, Obrazovaniya). 2009.
No. 7. P. 9–13. Available at: http://amnko.ru/index.php/english/journals (accessed June 1,
2021).
[14] Kirsta Y.B. Spatial generalization of climatic characteristics in mountain areas // World
Sci. Cult. Educ. (Mir Nauki, Kul’tury, Obrazovaniya). 2011. No. 3. P. 330–337. Available at:
http://amnko.ru/index.php/english/journals (accessed June 1, 2021).
[15] Kirsta Y.B., Kirsta B.Y. The information-physical principle of evolutionary systems forma-
tion. System-analytical modelling of ecosystems. 2nd ed. Barnaul: Altai State University
Publishing House, 2014. 283 p. (In Russ.)
[16] Kirsta Y.B. System analysis and forecast of climate in the grain-producing zone of Russia
until 2020 // Proceedings of the All-Russian Conference Dedicated to the 90th Anniversary
of Academician M.I. Budyko, 10–11 June 2010. St. Petersburg: BBM, 2010. P. 34–36.
[17] Afonin A.N., Greene S.L., Dzyubenko N.I., Frolov A.N. (eds.). Interactive agricultural
ecological atlas of Russia and neighboring countries. Economic Plants and Their Diseases,
Pests and Weeds [Online]. 2008. Available at: http://www.agroatlas.ru (accessed June 1,
2021).
[18] Kirsta Y.B. The forecast of both climate and agroclimatic potential in Siberian Federal
Okrug till 2020 // Regional Environmental Issues. 2011. No. 3. P. 22–30.
470