The impact of global climate change on the development of agroecosystems in Russia

Research results and climate change forecast over a long-term period

Forecasts by the Intergovernmental Panel on Climate Change indicate that the 21st century will be a period of unprecedentedly rapid climate change, which will have a significant impact on many sectors of the economy and, first and foremost, on agriculture.

Studies of the main climate characteristics over the last 10-15 years have shown that this period is the warmest and most humid not only in 100, but in 1000 years. Analysis of empirical climate data, conducted at the Institute of Water Problems of the Russian Academy of Sciences, revealed a statistically significant trend in the changes of various climate indicators, such as the dates of freezing and ice breaking of the largest rivers in Russia, earlier arrival of spring, increased duration of the growing season in the Northern Hemisphere, etc.

Scientific literature also reflects other anomalies caused by the rapid rates of warming and cooling. The main ones are as follows:

• acceleration of climate aridization processes, frequency of droughts;

• increase in CO2 and CH4 content in the atmosphere;

• impact of the observed "explosive" warming on the bioclimatic potential of agricultural territories;

• decrease and variability of agroecosystem bioproductivity;

• decrease in the level of soil fertility, their degradation;

• decrease in the sustainability of agroecosystem development.

According to calculations and forecasts by the international group of experts, which includes scientists from Russia, the increase in global average air temperature between 1990 and 2100 could range from 1.5 to 5.8°C. Such warming has no precedent over the past 10,000 years.

At the spring 2008 joint meeting of the Council for Environmental Problems in the Agro-Industrial Complex, with the participation of the Bureau of the Division of Agriculture, the Bureau of the Division of Land Reclamation, Water and Forestry of the Russian Academy of Agricultural Sciences, scientists from the FGBU "VNIISKhM", GNU "Agrophysical Research Institute of the Russian Academy of Agricultural Sciences" (AFI), the State Committee for Hydrometeorology, FGBNU VNII "Raduga", the Main Geophysical Observatory (GTO), and others, as well as agro-industrial practitioners, the importance of the climate warming problem and its impact on agriculture was recognized, as well as the need to develop a unified strategy for adapting agriculture to global climate changes, taking into account the preservation of soil fertility and resource conservation under conditions of increasing man-made pollution of the biosphere.

The research topics and modern methods for their implementation have been outlined. To solve the stated problem in the "climate-soil-harvest" system, a research topic has been justified, aimed at developing scientific and methodological foundations for assessing trends in climate factor changes, their impact on the bioproductivity of agroecosystems in relation to soil indicators, and coordination with production cycles, as well as the establishment of trends in climate indicator changes, including the integrated indicator of heat and moisture availability (Ky), and consequently, air temperatures (t°C), precipitation (P, mm), and evaporation (E, mm). The calculation is carried out by year over a long-term series using improved calculation models from FGBNU VNII "Raduga" and specialized computer software.

For an operational assessment of the impact of observed natural environment changes on agriculture, a system of normative characteristics is required. The task of creating a monitoring system for the impact of climate change appears relevant in this regard.

Thus, at the GNU AFI, theoretical and experimental data, monitoring observations, as well as results of work from other scientific and production organizations have been systematized and generalized. The main tasks have been formulated, on the basis of which a comprehensive research program is being compiled.

Changes in agroclimate lead to changes in the structure of sown areas, the productivity potential of agroecosystems, and to an increase or decrease in the use of agrochemicals and their spatial redistribution.

The greenhouse effect (change in atmospheric composition) is due to the fact that, depending on the method and intensity of agricultural land use, the level of agricultural technology, and the regulation of the level of soil fertility, as they warm up, they become a source of CO2.

Zonal fertilizer efficiency is manifested in the fact that yield increases from the amount of fertilizer applied decrease as one moves from humid regions to arid climate zones. The climatic factor has a stronger influence than the soil factor.

The redistribution of agroclimatic resources leads to the expansion of areas for cultivating crops, their cultivars and hybrids, which in turn will require the additional use of fertilizers.

Due to the dynamic nature of weather conditions, agronomic and land reclamation practices change. The need for regulating water and heat balance parameters is distributed as follows:

Research into the complex issues caused by global climate change is widely reflected in literature. Most works are devoted to the study and calculation of the impact of global climate warming on the main indicators of the "climate-soil-yield" system.

How climate and yield are changing: century-long trends

Climate changes directly affect field work. A decrease in heat and moisture availability forces agronomists to shift sowing dates, change the schedule of agrotechnical measures, and revise irrigation regimes. Due to the vagaries of weather, the yield of spring cereals in southern steppe regions can fluctuate significantly.

• Fluctuations in spring cereal yield — from 0.5–1 to 2–2.5 t/ha
• Growth in climate-dependent yield — by 30%
• Security of a 2 t/ha yield — increased more than 5-fold
• Global surface temperature rise over the century — by 0.6 ± 0.2 °C

In the last decades of the 20th century, climate-dependent cereal yield in southern steppe regions increased by 30% compared to the middle of the last century. At the same time, the guaranteed achievement of 2 tons per hectare became more than 5 times more probable. However, warming is uneven and affects only the troposphere — the surface layer of air several kilometers high, whereas temperatures in the upper layers of the atmosphere are decreasing.

The dynamics of temperature anomalies over a century of observations shows a pronounced cyclical nature. The main phases of temperature changes were distributed as follows:

Modeling the century-long course of spring wheat yield under unlimited nutrient supply (for temperate climate conditions over the period 1891–1990) revealed clear patterns. Two periods of growth in climate-dependent yield were recorded: in 1910–1920 and since 1991. During these times, the most favorable combination of heat and moisture occurred in the fields.

Adaptation and forecasting: new tools for the agronomist

To minimize the risks of droughts, dust storms, or waterlogging, it is necessary to adapt land reclamation systems to new cycles. Reclamation remains the main management factor in the field, but its efficiency depends on the accuracy of calculations. Irrigation and drainage regimes must correspond to current changes in the water, heat, and radiation balance of crops.

Modern computer programs and mathematical models are used for the scientific and methodological justification of irrigation and forecasting the moisture coefficient (K_c). They process long-term data from meteorological stations for any periods. Based on these calculations, agronomists can plan the irrigation season more accurately.

New software complexes allow for the prompt resolution of the following tasks:

• calculating chronological and probabilistic series of any climatic indicators;
• determining the degree and direction of weather trends for specific locations of meteorological stations;
• forecasting climate change cycles to synchronize them with the growth stages of crops.

The use of predictive models allows for the timely implementation of adaptation techniques in fields. As a result, the development of environmental crises is prevented, and the decline in soil fertility and biological productivity of land is halted.

Based on the calculations performed and their assessment using meteorological stations in Russian regions, an information and methodological base has been created and preliminary results obtained. A regional approach was used to solve the stated task. In doing so, the necessity of distinguishing between the concepts of climate "variability" and "change," which affect yield in different ways, was taken into account.

For a preliminary and to a certain extent approximate assessment of changes in climatic indicators of air temperature and precipitation in the Central Federal District, calculations were performed based on 60 years of observations from meteorological stations in the region.

When constructing graphs, average monthly air temperatures and total precipitation for the period from April to October of each year (1945–2003) from the Kolomna meteorological station were used. An integral method of dynamic modeling of the initial indicators was applied.

Modern methods for selecting probabilistic irrigation rates for any agricultural crop are based on an analysis of a series of irrigation rate values for a given crop, calculated from meteorological observation series available in the area. It is assumed that the meteorological conditions observed in the past, which determined the variability of irrigation rates by year, are an unambiguous forecast for the future.

Meanwhile, such an approach contradicts the probabilistic nature of the irrigation rate, as the meteorological situation over the available observation period represents a random process that may never repeat itself. Moreover, the analysis of meteorological observations indicates the possibility of the appearance of new, previously unrecorded, extreme phenomena. All this points to the fact that, in general, climate variability is such that the available period of regular meteorological observations does not allow for calculating stable values of the required statistical characteristics.

In other words, a series of irrigation norms calculated from available meteorological observation series, especially in the Non-Chernozem zone of Russia, where climate variability from year to year is very high, is not sufficiently representative, i.e., it does not cover the entire series or general population. Therefore, the traditional approach to modeling random sequences using statistical tests (the Monte Carlo method) does not always provide the desired result, since the model series must have the same statistical characteristics as the initial sequence.

The applied methods of constructing an approximating probability curve according to Foster-Rybkin based on a three-parameter gamma distribution, or according to Pearson type III distribution curves and their modifications by Kritsky-Menkel, are often not suitable for an adequate description of the empirical distribution, especially for extrapolation in the range of high values of the irrigation norm.

Searches for ways to increase the reliability of constructing probability curves under these conditions led to the conclusion that it is necessary to pre-process empirical distributions in order to identify the main patterns of irrigation norm distribution, which seem to be obscured by random variations in the values of irrigation norm frequencies from one interval to another. This conclusion is based on the premise that nature, acting randomly, cannot have preferred states that cause high values of irrigation norm frequencies in some intervals and low values in neighboring intervals. Therefore, over many years, nature must lead to a relatively uniform filling of intervals with a smooth, consistent transition of frequency values from one interval to another.

This premise is the basis for the approximating analytical curve that describes the approximated empirical distribution more or less accurately. Methodology for calculating Fourier series Let a function f(x) be defined on the interval [-1; 1] and its period be 2l. Then the following equalities hold: г/ \ /о / 7TYIX. 7TYIX /1 jl\ f(x)=ao/2+ y (ancos----- \-bnsin-----), (1ф) n=\ l l j a„=(l/l) f (x)cos— dx (n=0,l,2,...), (2ф)

b n = (l/l)j/(x)sin — dx (n = l,2,..). (ЗФ)

Series (1ф) with coefficients calculated by formulas (2ф), (ЗФ) is called the Fourier series for the function f(x) with period T=2l.

Let y=f(x) now be a non-periodic function (for example, f(x) is the average value of temperature or the sum of precipitation for the period April-October, x is the current year in this case). Such a function cannot be expanded into a Fourier series directly; however, it can be represented as a Fourier series on any finite interval [a, b] (in our case, a and b are the start and end years of the chronological series under consideration). To do this, it is necessary to place the origin of coordinates in the middle of the interval [a, b] and construct a function fi(x) of period T=2l=|b-a| such that fi(x)=f(x) for -l

In the example under consideration, for an approximate calculation of the function f(x), it is sufficient to calculate the sum of the first terms of the Fourier series: f(x)=ao/2+aiCos(7CxA)+bisin(7cx/l)+a2COs(27Cx/l)+b2sin(27ix/l), where ao/2 is the average value of temperature or the sum of precipitation for the period April-October;

l is half of the period under consideration |b — a|; x (due to the shift of the origin of coordinates) corresponds to the current year - |b — a| /2 - start year.

Coefficients a1, a2, b1, b2 are calculated using formulas (2ф), (Зф), in which the integrals are replaced by finite sums calculated on the interval [a, b] (dx = Δx = 1 year).

The trend of climate factor changes, calculated using Fourier series for the Kolomna weather station, is presented in fig. 6.

Data analysis based on the construction of spline functions for forecasting

Mathematical calculation of climatic cycles for irrigation planning

For effective irrigation, it is important to understand in which phase of the natural-climatic cycle the field is located during a specific period. Computational models of long-term cyclicity help the agronomist predict yield dynamics and choose adequate reclamation measures in advance. This allows for the synchronization of the farm's production cycles with natural weather fluctuations.

A spline function is used as a basis for constructing forecast models. The method allows for identifying the mathematical dependence of subsequent values of the moisture coefficient on the previous ones. Based on the cycles obtained, the trend of climatic condition development for future periods is determined.

The dependence of the moisture coefficient (K_u) on time is calculated using the following formula:

F(t) = a₀ + a₁t + a₂t² + a_K1cos(tw₁) + a_K2cos(tw₂) + a_K3cos(tw₃) + b_K1sin(tw₁) + b_K2sin(tw₂) + b_K3sin(tw₃)

To evaluate the accuracy of the obtained spline function and meteorological data, the following parameters are analyzed:

• autocorrelation coefficients, which show the relationship between deviations of calculated indicators from actual ones over time;
• correlation coefficients between air temperature sums and precipitation volumes;
• root-mean-square deviation, as well as maximum and minimum values of the moisture coefficient with a probability of falling within the interval of at least 95%.

The table shows the calculated spline function coefficients for key climatic indicators based on long-term observation data from meteorological stations:

For practical calculations, a computer program is used that automatically selects the version of the spline function with the minimum deviation from the actual meteorological station data. This allows for constructing the most probable trend of soil heat and moisture availability.

• Accuracy of weather forecasts — 15–20% higher
• Correlation ratio — 0.72–0.83
• Probability of falling within the interval — at least 95%