Terrain Elevation as a Driver of Anthropocene Trends in the Runoff of Rivers: Insights from the East European Plain

Abstract

Relief is an important driver in the spatial differentiation of river runoff and its regime both in the mountains and on the plains, which is most evident in arid and semi-arid regions of the Earth’s land. Based on 22 small and medium-sized rivers of the zones of forest–steppe and steppe in the temperate climate zone of the eastern part of the East European (Russian) Plain, within the Middle Volga region and the eastern part of the Don River basin, the role of this factor in the spatiotemporal changes in various key runoff parameters (annual average runoff (Q), annual maximum runoff (Q_max), and annual minimum runoff for both cold (Q_min-CP) and warm (Q_min-WP) seasons) between two baseline climatic periods of the Anthropocene (1961–1990 and 1991–2020) is considered using the average elevation of river basin (H) as its quantitative indicator and statistical procedures of regression and correlation analysis. It is found that in the interperiod trends of the Anthropocene, the H factor was the leading (and statistically significant) cause of spatial variability in the changes in Q_max in the forest–steppe zone, through its inverse relationship with H (with a 70% contribution of influence), Q_min-CP in the forest–steppe and steppe zones (with a 55–75% contribution of influence), and Q_min-WP in the steppe zone (with a 64% contribution of influence), through their direct relationships with H. It is also shown that H acted as an important factor (with a 47% contribution of influence) of statistically significant strengthening of the spatiotemporal coherence of Q and Q_max values between the studied periods, but only in the river basins of the Middle Volga region: during the period of the most active climate warming (1991–2020), the region’s upland rivers turned out to be more coherent in these two runoffs than the lowland ones. An ambiguous influence of H on the mutual correlation of all the examined runoff parameters over the baseline periods was revealed. The identified patterns are a consequence of the reaction of the complex altitudinal zoning of the plain’s landscapes mainly to climate changes, especially during the cold season (frequent thaws, decreasing soil freezing depth, etc.). The achieved results are intended to contribute to the understanding of the role of so-called “passive” driving forces in contemporary regional changes in river runoff.

1. Introduction

In the context of ongoing progressive climate change, the study of spatiotemporal variability in freshwater resources and its factor assessment are one of the priority scientific and applied issues. This article is a conceptual, territorial and factual continuation of our previous paper [1], dedicated to identifying the prevailing trends in the change in the runoff of water (hereinafter, just runoff) of small and medium-sized rivers in the forest–steppe and steppe landscape environments of the East European (or Russian) Plain’s eastern part during the Anthropocene. The Anthropocene, as a current geological epoch, is characterized by the fact that human activity, since its beginning, has become one of the leading forces influencing many natural processes on our planet [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16]. The lower boundary (the beginning) of the Anthropocene (the Great Acceleration) is often considered to be the dropping of the first atomic bombs in 1945, and especially nuclear weapons testing in the 1950s–1960s, which left a global trace of radioactive isotopes (e.g., ²³⁹Pu and its derivatives) that will be detectable in the environment for thousands of years to come [17,18,19]. Although the term “Anthropocene” has not yet been officially approved by the International Union of Geological Sciences, it is widely used in science.

The climatic, hydrological and land use/cover contexts in which the Anthropocene changes in runoff occurred in the study area, as well as the reasons for these changes, are also presented in the same work [1]. In general terms, these contexts are as follows.

(1)

The warming of the lower troposphere within Russia is occurring almost twice as fast as the average warming on the surface of the Earth’s land—approx. 0.5 °C per 10 years; moreover, each subsequent decade, starting from the 1981–1990 period, was warmer than the previous one [20]. Along with the observed increases in air temperature, the last decades have also seen a decline in the duration of snow cover, especially in European Russia (across Russia as a whole, an average of 1.2 days over 10 years from 1976 to 2020). However, the total snowfall amount has increased, particularly in the western part of Russia [21]. In the southern half of European Russia, especially in the steppe zone, there is a slight increase in the amount of heavy rainfall during the warm season of the year [22].

(2)

Following the USSR’s collapse in 1991, there was a considerable reduction in arable land (which has sharply slowed or even reversed since the 2010s), especially in the region’s forest zone, as well as changes in crop rotation patterns [23,24,25,26]. Based on our estimates [27], the area of cropland in Russia as a whole decreased by more than a third between 1971–1991 and 2005–2017.

(3)

Climate and land use/cover changes have led to considerable seasonal transformations in runoff, primarily in the southern half of the plain. This has resulted in a reduction in meltwater runoff and an increase in low-water runoff during both cold and warm seasons of the year. For example, in the basins of the lower reaches of the Volga (as the longest and most full-flowing river in Europe) and Don rivers, as well as in the southern part of the basin of the Oka River, a decrease in the annual maximum runoff of 40–70% was recorded in most rivers after 1979–1983 [1,28,29,30]. In the basin of the Don River and in the eastern part of the basin of the Volga River, the contribution of floodwaters from snowmelt has decreased to ≤50% of the annual runoff [30,31,32] (50% is a critical level for rivers with an Eastern European-type water regime). In the first half of the 20th century, for comparison, snowmelt floodwaters in these basins accounted for ≥60–70% of annual values [29,30].

A fairly large number of works are devoted to the analysis of climatic and, to a lesser extent, anthropogenic causes of the above-mentioned changes in runoff in European Russia [20,21,22,27,28,29,30,31,32], and their list continues to expand. Despite this, the Russian scientific community has paid undeservedly little attention to the influence of more conservative (less dynamic) factors on modern trends in river runoff changes, which often act as a kind of driver of these changes. This primarily concerns geological (lithology, tectonic structure, etc.) and geomorphological regional conditions, the influence of which is often complementary. The role of relief, for example, is manifested in its influence on runoff intensity and its spatiotemporal dynamics in various ways, primarily through its topographic characteristics (absolute and relative elevation, slope and aspect, morphology, etc.).

The East European Plain, most of which comprises European Russia, is an alternation of lowlands and uplands. The densest alternation of uplands and lowlands is found in the plain’s southern half, within the forest (its southern sub-belt), forest–steppe and steppe landscape zones [33]. These uplands and lowlands are the East European Plain’s largest orographic elements; therefore, they are the principal regulating factors of the climatically determined spatial distribution of runoff. This primarily concerns the terrain elevation, which directly and indirectly regulates the spatiotemporal variability of runoff by regulating the fallout of liquid and solid (snow) precipitation and its evaporation (through the exposure of the slopes, absolute elevation, etc.).

The aim of this study is to identify the impact of terrain elevation using the average basin elevation (H) as the examined factor on the Anthropocene changes in the runoff of small and medium-sized rivers of the southern half of the East European Plain, within its steppe and forest–steppe landscape–climatic zones, by solving the following paramount tasks:

• To identify the nature of the impact of H on the changeability of the annual average, maximum and minimum (for the cold and warm seasons of the year) runoff both within the framework of the baseline periods, 1961–1990 and 1991–2020 (spatial variability), and between these periods (spatiotemporal variability).
• To detect spatial variability of the runoff coherence (i.e., manifestation of synchronicity of temporal variability in runoff) of the above-mentioned runoff parameters both within the specified baseline periods and between these periods, and to assess the influence of H on this coherent variability within two subregions of the region under study, the Middle Volga region and the eastern Don River basin.
• To assess the influence of H on the change in the direction and closeness of the mutual correlation between the above-mentioned runoff parameters from 1961–1990 to 1991–2020.

Thus, the results of the study present the first systematic quantitative assessment of the contribution of relief elevation to contemporary (Anthropocene) trends in various key parameters of small and medium-sized river runoff and their spatiotemporal coherence in the study region. They are intended to provide only a general and preliminary understanding of the multidimensional role that relief plays in modern hydrological processes and their geography.

2. Materials and Methods

2.1. Study Region

The region under study is located in the eastern part of the East European Plain and encompasses such major orographic divisions as the Central Russian Upland (its eastern megaslope), Volga (or Cis-Volga) Upland, Bugulma–Belebey Upland (its western megaslope), Obshchy Syrt Upland (its western megaslope), and the lowlands between them: Oka–Don Lowland, Trans-Volga Lowland, Caspian Lowland, etc. (Figure 1). Sedimentary cover rocks, which vary in ages (from Late Paleozoic to Quaternary deposits) and lithology, compose these divisions [33,34].

Figure 1. The region under study and analyzed rivers (according to our previous work [1] with some changes). Labels 1, 2, …, and 22—the numbering of the rivers (Table 1); I and II—Middle Volga region and the eastern Don River basin, respectively; BBU—Bugulma–Belebey Upland; OSU—Obshchy Syrt Upland; ODL—Oka–Don Lowland. A—the studied gauging stations on the rivers; B—some largest cities; C—seas, rivers, reservoirs, and lakes; D—forests (except for the Caspian and Aral lowlands); E—cultivated land; F—meadows and pastureland; G—the border between the Volga River and Don River basins within the study region.

Figure 1. The region under study and analyzed rivers (according to our previous work [1] with some changes). Labels 1, 2, …, and 22—the numbering of the rivers (Table 1); I and II—Middle Volga region and the eastern Don River basin, respectively; BBU—Bugulma–Belebey Upland; OSU—Obshchy Syrt Upland; ODL—Oka–Don Lowland. A—the studied gauging stations on the rivers; B—some largest cities; C—seas, rivers, reservoirs, and lakes; D—forests (except for the Caspian and Aral lowlands); E—cultivated land; F—meadows and pastureland; G—the border between the Volga River and Don River basins within the study region.

Table 1. The analyzed rivers and some characteristics of their basins (see Figure 1), according to our previous work [1].

No.	River	Gauging Station and Its Code 1		F, km2	h, m	H, m	Z	Lith
Forest–steppe zone
1	Myosha	Pestretsy	77197	3230	54.9	150	5.4	C
2	Kubnya	Chuteyevo	77164	930	77.3	153	3.8	C
3	Kichui	Utyashkino	77189	1330	51.0	185	5.3	C
4	Aktai	Karavayevo	77201	690	69.7	132	3.1	L
5	Sheshma	Sloboda Petropavlovskaya	77179	3110	59.8	205	5.0	C
6	Maliy Cheremshan	Abalduyevka	77217	1230	85.3	144	3.7	L
7	Krasnaya	Krasnaya Reka	77209	311	55.0	120	2.9	L
8	Bol’shoy Cheremshan	Novocheremshansk	77212	6050	59.5	143	3.5	C
9	Tushonka	Sergeyevka	77210	309	54.2	227	4.0	C
10	Syzranka	Repyovka	77329	4380	48.0	220	3.4	Sed
11	Sosna	Yelets	78054	16,300	106.9	210	3.7	S-S
12	Vorona	Borisoglebsk	78165	13,200	84.5	ND	3.0	Ch-O
				4256	67.2	172	3.9
Steppe zone
13	Maliy Kinel’	Poludni	77298	2090	49.5	165	2.6	C
14	Bol’shoy Kinel’	Timashevo	77292	12,000	32.2	165	3.2	C
15	Chapayevka	Pod’yom Mikhailovka	77311	1480	48.0	132	1.7	Lo
16	Chagra	Novotulka	77336	2550	29.7	107	1.2	L
17	Buzuluk	Perevoznikovo	77270	4280	61.7	162	1.6	C
18	Khopyor	Balashov	78138	14,300	100.8	220	3.1	Ch-O
19	Medveditsa	Lysyye Gory	78196	7610	126.6	220	2.6	C
20	Bol’shoy Karaman	Sovetskoye	77362	3470	29.0	82	0.4	L
21	Ilovlya	Borovki	78231	8730	40.3	150	0.8	S-S
22	Chir	Oblivskaya	78252	8470	39.2	150	1.1	Lo
				6498	55.7	155	1.8

Notes: ¹ According to https://gmvo.skniivh.ru (accessed on 20 February 2025); F—river basin area upstream of the analyzed gauging station; h—the zero level at the analyzed gauging station (m a.m.s.l.); H—the average absolute river basin elevation upstream of the analyzed gauging station; Z—the long-term mean annual average specific runoff in 1961–2022, L s⁻¹ km⁻²; Lith—surface deposits’ lithologic complexes: L—loams, Lo—loess and loess-like loams, C—clays, S-S—sand–shale rocks, Ch-O—chemogenic and organogenic rocks, and Sed—undifferentiated sedimentary deposits. The average values of the characteristics in each landscape zone are in bold. ND—no data. Note: Data on H and Lith are from [35].

Table 1. The analyzed rivers and some characteristics of their basins (see Figure 1), according to our previous work [1].

No.	River	Gauging Station and Its Code 1		F, km2	h, m	H, m	Z	Lith
Forest–steppe zone
1	Myosha	Pestretsy	77197	3230	54.9	150	5.4	C
2	Kubnya	Chuteyevo	77164	930	77.3	153	3.8	C
3	Kichui	Utyashkino	77189	1330	51.0	185	5.3	C
4	Aktai	Karavayevo	77201	690	69.7	132	3.1	L
5	Sheshma	Sloboda Petropavlovskaya	77179	3110	59.8	205	5.0	C
6	Maliy Cheremshan	Abalduyevka	77217	1230	85.3	144	3.7	L
7	Krasnaya	Krasnaya Reka	77209	311	55.0	120	2.9	L
8	Bol’shoy Cheremshan	Novocheremshansk	77212	6050	59.5	143	3.5	C
9	Tushonka	Sergeyevka	77210	309	54.2	227	4.0	C
10	Syzranka	Repyovka	77329	4380	48.0	220	3.4	Sed
11	Sosna	Yelets	78054	16,300	106.9	210	3.7	S-S
12	Vorona	Borisoglebsk	78165	13,200	84.5	ND	3.0	Ch-O
				4256	67.2	172	3.9
Steppe zone
13	Maliy Kinel’	Poludni	77298	2090	49.5	165	2.6	C
14	Bol’shoy Kinel’	Timashevo	77292	12,000	32.2	165	3.2	C
15	Chapayevka	Pod’yom Mikhailovka	77311	1480	48.0	132	1.7	Lo
16	Chagra	Novotulka	77336	2550	29.7	107	1.2	L
17	Buzuluk	Perevoznikovo	77270	4280	61.7	162	1.6	C
18	Khopyor	Balashov	78138	14,300	100.8	220	3.1	Ch-O
19	Medveditsa	Lysyye Gory	78196	7610	126.6	220	2.6	C
20	Bol’shoy Karaman	Sovetskoye	77362	3470	29.0	82	0.4	L
21	Ilovlya	Borovki	78231	8730	40.3	150	0.8	S-S
22	Chir	Oblivskaya	78252	8470	39.2	150	1.1	Lo
				6498	55.7	155	1.8

Notes: ¹ According to https://gmvo.skniivh.ru (accessed on 20 February 2025); F—river basin area upstream of the analyzed gauging station; h—the zero level at the analyzed gauging station (m a.m.s.l.); H—the average absolute river basin elevation upstream of the analyzed gauging station; Z—the long-term mean annual average specific runoff in 1961–2022, L s⁻¹ km⁻²; Lith—surface deposits’ lithologic complexes: L—loams, Lo—loess and loess-like loams, C—clays, S-S—sand–shale rocks, Ch-O—chemogenic and organogenic rocks, and Sed—undifferentiated sedimentary deposits. The average values of the characteristics in each landscape zone are in bold. ND—no data. Note: Data on H and Lith are from [35].

One of the most extensive uplands of the plain, the Central Russian Upland (about 480,000 km²), is located in its central part. The upland’s highest elevation reaches 310 m. Unlike other uplands within the plain, it lacks sharp elevation fluctuations, most likely due to the weak differentiation of recent tectonic movements [33,34]. Erosional dissection by river valleys and small dry valleys is very high. The vast Volga Upland (about 800 km long and about 500 km wide) stretches along the Volga River’s right bank. The average absolute elevation of the upland is about 200 m, with a maximum of 379 m. The Volga Upland is characterized by a dense network of small dry valleys and gullies—up to 1.9 km per km² [33,34]. The High Trans-Volga region consists of the Bugulma–Belebey Upland (Figure A1 in Appendix A), where relative elevation fluctuations reach 200–300 m, and the Obshchy Syrt Upland (up to 405 m) [33,34]. The gently undulating surfaces of these uplands are dissected by deeply incised, asymmetrical ancient river valleys and younger small dry valleys.

The Oka–Don Lowland is located between the Central Russian and Volga uplands. It is extensive, with absolute elevations of 150–180 m (in the south, up to 55 m) [33,34]. The Trans-Volga Lowland extends sub-meridionally along the Volga River’s left bank, from the city of Kazan in the north (approx. 20 km wide) to near the city of Volgograd in the south (up to 250–300 km wide) (see Figure 1). On its eastern edge, the lowland gradually transitions into the elevated High Trans-Volga region, and in the south, it merges with the Caspian Lowland. At the border with the Caspian Lowland, absolute elevations are close to zero, in the north—125–150 m, and in the interfluves—up to 200 m [33,34]. The surface of the lowland is composed of Quaternary sediments.

The region under consideration is located in the forest–steppe and steppe landscape zones with a moderate continental climate (Dfb, according to the Köppen climate classification) and a continental climate (Dfa) with increasing general continentality towards the east and south–east. In the forest–steppe, summers are warm (in July, an average of +20° to +22 °C), and winters are noticeably colder and harsher than in the west of the plain (in January, the average air temperatures reach up to −15 °C) [33,34]. Precipitation is 400–500 mm per annum. In the steppe region, summers are hot (in July, +22° to +24 °C), and winters are cold (in January, the average air temperatures reach up to −16 °C). Precipitation is 250–400 mm per annum, and potential evaporation significantly exceeds precipitation [33,34].

The average long-term runoff depth within the forest–steppe is approx. 110 mm per annum. This is significantly less than in the taiga zone located to the north (300 mm per annum and more), but significantly greater than in the steppe, where it is only 40–50 mm per annum. The distribution of runoff by seasons is highly inhomogeneous. The share of spring snowmelt-induced runoff in the forest–steppe is approx. 60–65% (it has appreciably decreased in recent decades). Summer–fall and winter low-water periods are clearly expressed. In the steppe zone, the intra-annual distribution of runoff is extremely inhomogeneous. From 70% to 90% of the annual runoff (with the maximum in the eastern part of the zone) occurs during the short spring snowmelt flood. Water supply to the rivers is mixed, still with a predominance of snowmelt water.

The region’s preponderant soils are gray forest soils under forests (Greyzems, or Gray Luvisols) and Chernozems (leached, podzolized, deep, etc.) under the grassy steppe vegetation. Greyzems, and especially Chernozems, are plowed over a significant part of their distribution [33,34]. Forest–steppe vegetation consists of alternating forested areas and patches of steppe communities. Steppe vegetation consists of forb–grass species and is now mostly plowed.

2.2. Analyzed Rivers

The study embraces 22 small and medium-sized rivers from the Volga River and Don River basins, between 48°28′ N and 56°11′ N and between 36°21′ E and 54°06′ E (see Figure 1). The total area of their basins (upstream of the gauging stations under consideration) is 116,050 km² (51,070 km² in the landscape zone of forest–steppe (44%) and 64,980 km² in the zone of steppe (56%)). Information on the elevational, hydrological, and lithologic characteristics of these rivers’ basins is adduced in Table 1. It should be noted that the average elevation of the steppe zone’s basins is slightly lower than in the forest–steppe zone due to the overall decrease in elevation in the southeastern part of the plain under consideration towards the Caspian Lowland (Depression). The long-term mean annual average specific runoff (Z) in the forest–steppe basins considered is more than twice as intense as in the considered steppe zone basins (see Table 1).

Spatial variability in runoff in small and medium-sized river basins is closely related to the geological and geomorphic features of the study region, which is especially important as the climate becomes increasingly continental: in steppe river basins, the hypsometric factor manifests itself more (statistically) significantly than in forest–steppe basins (Figure 2). This is particularly true for river basins located at relatively higher hypsometric levels and composed of less permeable surface rocks (clay, chemo- and organogenic ancient deposits, etc.). Conversely, in river basins within lowland plains, composed of comparatively looser sediments, the annual average specific runoff (Z) is, on average, the lowest (Figure 2D). According to the relative position of the trend lines in Figure 2C, with equal average elevations of river basins (H), Z is higher in the zone of forest–steppe than in the zone of steppe. Moreover, this difference is greatest within lower-lying river basins (at H ≤ 150 m a.m.s.l.; in the forest–steppe zone, we do not have data on the runoff in basins with H < 120 m a.m.s.l.), and decreases as H increases (especially at H ≥ 210 m a.m.s.l.).

2.3. Materials

The study materials are presented by long-term observation series (from 1961 to 2020) of water discharge (m³ s⁻¹) in the analyzed rivers according to the following key runoff parameters:

▪

Q—Annual average runoff;

▪

Q_max—Annual maximum runoff (during the spring flood caused by snowmelt);

▪

Q_min-CP—Annual minimum runoff during the ice-covered riverbed period (hereinafter, referred to as the cold period (season), mostly December–March);

▪ Q_min-WP—Annual minimum runoff during the absence of ice cover in the river (hereinafter, referred to as the warm period (season), mostly April–November).

For all the rivers under consideration, long-term series of all the specified runoff parameters were considered, excluding the Maliy Kinel’ River, for which only data on Q and Q_max were available. All original observation series are presented in detail in our previous article [1]. Information on filling in the gaps in these series can also be found in the mentioned article. The initial data for the above runoff parameters were derived from the automated information system for state monitoring of water bodies of the Federal Agency for Water Resources of the Ministry of Natural Resources and Environment of Russia (https://gmvo.skniivh.ru/, accessed on 17 January 2025), reference hydrological sourcebooks (see [1]), and others.

2.4. Methods

2.4.1. Analyzed Periods

The analysis of runoff changes in the context of terrain elevation changes was carried out for two periods: 1961–1990 and 1991–2020. These periods were chosen because they serve as reference periods for long-term climate change observations on a global scale: 1991–2020 (P₁) is the updated 30-year period of a new climatological normal, and 1961–1990 (P₂) represents the previous baseline period. Both of these periods are highlighted according to the US National Oceanic and Atmospheric Administration. The period 1961–1990 is still recommended by the World Meteorological Organization for historical comparisons of climate change (P₁ vs. P₂). Moreover, our previous study [1] showed that the period from roughly the late 1980s or the early 1990s was an important time boundary for statistically significant changes in low-water runoff in the study region and, consequently, in intra-annual variability in annual runoff as a whole. Furthermore, as mentioned above, this time boundary coincided with another important regional milestone, namely an important shift in regional land use (land cover), caused by the Soviet Union’s collapse in 1991 [1,23,24,25,26,27]. Thus, the choice of the indicated periods was dictated by a combination of circumstances of both global and regional significance.

2.4.2. Statistical Procedures

In this study, the paramount methods were regression and correlation analysis, and Spearman’s correlation coefficient (r) was used as a tool for the latter. This coefficient was chosen due to the following advantages: it evaluates the monotonic relationship, which is a broader category of evaluation; low sensitivity to outliers; it is a robust measure of relationship, which provides a more reliable estimate if the data contains suspicious values or “tails” of the distribution that cannot be discarded; and it is a nonparametric criterion, which does not require that the data obey any specific distribution law (the latter circumstance makes it a universal tool for exploratory analysis and working with real data, which are rarely perfectly normal).

The following additional statistical procedures were also used:

• The Mann–Kendall test for the presence of a trend and its statistical significance (p-values). The trend line approximation coefficient R² was also considered as a determination coefficient showing the contribution (in %) of a changing factor (in our case, this is the terrain elevation variability) to the variability of the dependent value (the variability of runoff).
• A set of tests for checking the homogeneity of the obtained correlation relationships (Buishand’s test, Pettitt’s test, and the Standard normal homogeneity test) to identify critical elevation levels of runoff change (the elevation level that matched the results of at least two of the three proposed tests was considered critical).
• White’s test for checking the heteroscedasticity or homoscedasticity of the achieved correlation relationships, i.e., the presence or absence of a unit root in the residuals of the correlation relationship trend, respectively.

2.4.3. Analytical Stages of the Study

(1)

Terrain Elevation and Spatiotemporal Variability of Runoff. Correlation of the average elevations of the river basins was analyzed (H) with average values of considered runoff parameters in the baseline climatic periods in the forest–steppe and steppe landscape zones of the East European Plain. In this correlation, all average values of a particular runoff parameter (m³ s⁻¹; Q, Q_max, Q_min-CP, and Q_min-WP) were converted into specific values (L s⁻¹ km⁻²; respectively, Z, Z_max, Z_min-CP, and Z_min-WP) in order to avoid the influence of the river basin (river length) size factor on the homogeneity of the achieved relationships. The correlation trend between H and the corresponding runoff parameter was considered statistically significant at p ≤ 0.05. This correlation was implemented computationally and graphically in the XLSTAT 2016.02.28451 statistical package (for Microsoft Excel).

(2)

Terrain Elevation and Spatiotemporal Coherence of Runoff. In each baseline climatic period, for the corresponding runoff parameter, a mutual correlation was carried out for all rivers included in the two study subregions: the Middle Volga region and the eastern part of the basin of the Don River as the most compact subregions of the location of the analyzed river basins in the region. In each subregion, for each river, individual mutual correlation coefficients (r_av) were computed, which were then used to calculate the average correlation coefficients for the entire subregion by period (r′). The higher the values of r_av and r′, the higher the spatiotemporal coherence of the change in the runoff parameter for the corresponding river (relative to other rivers in the subregion) and the subregion as a whole, respectively (complete coherence of the runoff in the river (or in the subregion) is observed at r_av (or r′) = 1.0). The constructed mutual correlation matrices also make it possible to identify both the most representative (with the highest individual correlation coefficients) and “anomalous” rivers (with extremely low coefficients) for further study of the behavior of the long-term dynamics of their runoff. The interbasin mutual correlation procedure was implemented computationally and graphically in the OriginPro 2025b SR1 installation package.

(3)

Terrain Elevation and Intra-annual Runoff Correlation. For each river analyzed, a mutual correlation was carried out between the corresponding runoff parameters in each of the baseline periods. Matrices of such correlations make it possible to identify both the direction and the closeness and statistical significance of intra-annual correlation relationships between the parameters. An example of such a matrix is presented in Figure A2 in Appendix A, using data on the runoff of the “anomalous” Krasnaya River, which flows in the Middle Volga region’s forest–steppe zone. The derived results were then also generalized for the landscape zones under consideration. The mutual correlation procedure was also implemented computationally and graphically in the OriginPro 2025b SR1 installation package.

2.5. Limitations

The following principal limitations of this work should be noted:

(1)

There is a relatively small number of analyzed rivers in the Don River basin. This makes it difficult, on the one hand, to identify statistically reliable and stable dependencies of trends in runoff change on the elevation of the terrain in the river basin itself, and on the other hand, to reveal statistically significant similarities and differences with the neighboring Middle Volga region.

(2)

The average elevations of river basins, although they are a very convincing indicator of intraregional differences in topography, do not give an idea of the distribution of elevations, especially relative ones, within the river basins themselves without knowledge of the average slope gradients (α) in them. Considering the combined impact of these two geomorphic factors (H and α) on runoff will contribute to a more successful identification of stable geomorphic–hydrological dependencies in the region and their use in both scientific and applied issues.

(3)

Human activity in river basins (land use/cover changes, water intake for irrigation, creating ponds, etc.) could have had some impact on the identified runoff trends, as well as on interannual and intra-annual runoff variability during the baseline periods under consideration. We did not take this activity into account in our study. This issue requires separate and careful consideration within the framework of a single multifactorial analysis of the topic under study.

3. Study Background

In our previous study [1], the following main patterns of change in the river runoff trend in the forest–steppe and steppe landscape zones of the study area were identified:

• Between 1960 and 2022, the vast majority of the rivers analyzed (amounting to 77% across the two zones combined) showed an overall downward trend in Q. It is important to note that, despite the presence of prevailing downward trends in Q across the entire study region, in 82–86% of cases, these trends were statistically insignificant.
• Unlike the annual average runoff, Q_max showed a downward trend in both overall and seasonal levels across both considered zones. This decline was more noticeable and generally statistically significant. It is noteworthy that the decline in Q_max in the analyzed rivers in the zone of forest–steppe was, on average, “deeper” (with statistically significant differences) compared to the steppe zone’s rivers.
• Q_min-CP experienced a significant overall increase from 1961 to 2022 (by an average of 96–100% between the periods 1961–1990 and 1991–2020). Most of these trends were statistically significant, particularly in the forest–steppe zone.
• As with Q_min-CP, Q_min-WP had a predominantly increasing trend during the total period studied: in 92–100% of cases in the forest–steppe zone and 78% of cases in the steppe zone.
• The above-mentioned alterations in the examined runoff parameters also entailed changes in their ratios. For example, the decreases in the Q_max/Q_min-CP and Q_max/Q_min-WP ratios between 1961–1990 and 1991–2020 were, on average, 70–73% in the analyzed forest–steppe zone rivers and 65–68% in the steppe zone.
• In the study area as a whole, a slight reduction (by 5–6%) in the interannual changeability (variation coefficient) of Q was noted between 1961–1990 and 1991–2020.

The noted trends are well illustrated by the examples of the Bol’shoy Cheremshan and Khopyor rivers (Figure 3) as the most representative for the Middle Volga region and the eastern Don River basin, respectively, according to [1] (also see below).

4. Results

4.1. Terrain Elevation and Spatiotemporal Variability of Runoff

The role of relief regulating runoff was most clearly manifested in the steppe zone’s analyzed river basins rather than in the forest–steppe ones. In the basins of the steppe zone, the influence of H increased more in the baseline period 1991–2020 compared to 1961–1990. If in 1961–1990 the contribution of terrain elevation to the spatial variability of the annual specific runoff (Z) was estimated at 53%, according to the value of R², then in 1991–2020 it increased, while the mean zonal runoff values remained almost unchanged between the two periods (Figure 4). Positive linear trends, as the most optimal models of relationships, Z = f(H), in this case, were statistically significant and very close, according to the high correlation coefficients (r ≥ 0.7). The used tests of Z-series homogeneity showed that the critical H value for steppe rivers was 162 m a.m.s.l.: if in the period 1961–1990 the mean Z value in upland rivers (H > 162 m) was almost 2.3 times (127%) higher than in lowland rivers (H ≤ 162 m), then in the period 1991–2020 this difference was even greater, almost 2.8 times (182%).

Nothing similar was observed in the region’s more humid forest–steppe zone, except that between the baseline climatic periods, the annual average specific runoff values also remained nearly unchanged. A slight increase (a positive correlation) in the influence of H was noted only during the baseline period 1991–2020, but it was not statistically significant (see Figure 4).

Contrary to the annual average specific runoff, the variability of the annual maximum specific runoff (Z_max) showed a relationship with H (a negative correlation) only in the forest–steppe zone (Figure 5). This relationship was inverse, close, and statistically significant. Moreover, the relationship tended to weaken from 1961–1990 to 1991–2020, while the mean zonal values of this runoff also decreased. The tests used for Z_max series homogeneity showed that the critical H value for the analyzed forest–steppe rivers was 153 m: during both periods, the average Z_max values in upland rivers (H > 153 m) were almost 2.1 times (52–53%) lower than in lowland rivers (H ≤ 153 m).

In the steppe zone’s rivers, terrain elevation as a driver of Z_max did not manifest itself in any way in both studied periods, despite the long-term mean annual maximum specific runoff between them decreased by almost a third. In contrast to Z and Z_max, the minimum specific runoff in the cold season (Z_min-CP) largely depended (a positive correlation) on the regional relief in both landscape zones (Figure 6). The contribution of H to the overall spatial variability of Z_min-CP was at least 50% of the total set of controlling factors. Moreover, the influence of the relief factor H was more noticeable in the second baseline period, 1991–2020, than in 1961–1990 (see Figure 6). High values of the correlation coefficient (r > 0.7) in all cases indicate a close relationship, Z_min-CP = f(H), against the background of an overall increase in mean Z_min-CP values between the periods.

As in previous cases, the same critical values of H served as the boundaries of statistically significant changes in Z_min-CP: 153 m in the forest–steppe zone and 162 m in the of steppe zone. If in the forest–steppe zone the excess of average Z_min-CP values for upland rivers (H > 153 m) over lowland rivers (H ≤ 153 m) decreased from 2.7 times (or 174%) in 1961–1990 to 2.3 times (or 129%) in 1991–2020, then in the steppe zone it was considerably greater and increased from 5.6 times (or 462%) in 1961–1990 to 6.3 times (or 532%) in 1991–2020, with a threshold H value of 162 m. All of the differences reported were statistically significant.

As with Z_min-CP, the relationships between the changeability of the minimum specific runoff of the warm season (Z_min-WP) with changes in H were positive and relatively close, but in most cases statistically insignificant (excluding the analyzed steppe rivers in 1991–2020), although the contribution of the relief factor H to this variability ranged from 33% to 50% among other factors (Figure 7).

The same critical values of H also served as the boundaries of statistically significant changes in Z_min-WP: 153 m in the forest–steppe zone and 162 m in the steppe zone. If in the forest–steppe zone the excess of average Z_min-WP values for upland rivers (H > 153 m) over lowland rivers (H ≤ 153 m) decreased from 3.0 times (or 197%) in 1961–1990 to 2.6 times (or 164%) in 1991–2020, then in the steppe zone it was considerably greater—6.3 times (or 525%) both in 1961–1990 and 1991–2020—with a threshold H value of 162 m. Although the relationships, Z_min-WP = f(H), were not statistically significant in most cases, all differences identified across elevation groups showed statistical significance.

In all the above-considered cases of relationships Z = f(H), Z_max = f(H), Z_min-CP = f(H), and Z_min-WP = f(H), the most optimal form of trend approximation was linear among the most probable. Its optimality was determined by the highest value of R².

As for the difference in the values of the examined runoff parameters of each river averaged over the baseline periods (ΔZ), it was also influenced by changes in the average elevation of the basins. In changes in ΔZ of the annual average specific runoff, the influence of the relief factor H was statistically significant, but relatively small (no more than a third of the total influence of all factors of spatial variability of ΔZ values) only in the steppe zone’s rivers (Figure 8). Within the steppe river basins with average elevations >150 m, the average values of annual average specific runoff tended to increase from 1961–1991 to 1990–2020. In contrast, in lower-lying river basins (H ≤ 150 m), the annual average specific runoff tended to decrease.

The influence of the relief factor H was most clearly manifested in the reduction in annual maximum specific runoff in the forest–steppe zone (see Figure 8). The most noticeable reduction in runoff occurred in the baseline period 1991–2020 in lower-lying river basins (H ≤ 150 m), not in upper-lying river basins. At the same time, in the steppe zone’s river basins, a zero response of the annual maximum specific runoff to changes in absolute terrain elevations was observed.

As the average elevation of river basins increased, the long-term mean values of annual minimum runoff in both cold and warm seasons also increased in 1991–2020 compared to the previous 30-year baseline period (see Figure 8). This was especially evident in the steppe river basins: the contribution of the relief factor H to the indicated changes exceeded 60%. Despite the overall interperiod increase in the minimum specific runoff of the warm season with increasing terrain elevation in the forest–steppe zone, the contribution of the H factor was low and statistically insignificant.

To summarize the above, we can formulate the following:

(a)

Through the system of direct and inverse relationships, the considered relief factor H acted as the principal reason for the spatial variability of the annual average specific runoff within the river basins of the steppe zone in both baseline periods; the annual maximum specific runoff in the forest–steppe zone’s river basins in 1961–1990; and the minimum runoff of the cold season in both baseline periods and both landscape zones, except for 1961–1990 in the forest–steppe, where this relationship was not statistically significant (

Table 2. The contribution (the share, according to R² values) of the relief (the average elevation H) to intraperiod (based on data from Figure 4, Figure 5, Figure 6 and Figure 7) and interperiod (based on data from Figure 8) spatial variability of the mean values of the examined parameters of the specific runoff in the basins of the analyzed rivers.

Parameter	Landscape Zone	Intraperiod Contribution, %		Interperiod Contribution, %
		1961–1990	1991–2020
Q	Forest–steppe	2 ns	18 ns	31
	Steppe	53	64	64 ns
Qmax	Forest–steppe	59	40	70
	Steppe	0 ns	0 ns	0 ns
Qmin-CP	Forest–steppe	50 ns	61	55
	Steppe	55	63	75
Qmin-WP	Forest–steppe	47 ns	44 ns	18 ns
	Steppe	33 ns	50	64

Note: ns—not significant.

). With respect to the minimum runoff in the warm season, the positive influence of H was statistically significant only in the second baseline period in the steppe zone’s rivers, where its contribution amounted to half of the total contribution of the controlling factors (see Table 2).

(b)

In the interperiod trends of the Anthropocene, the analyzed relief factor H was the leading cause of spatial variability in the changes in the following parameter of specific runoff (ΔZ): Z_max in the forest–steppe zone, through the inverse relationships ΔZ = f(1/H); Z_min-CP in both zones, and Z_min-WP in the steppe zone through direct relationships, ΔZ = f(H) (see Table 2).

4.2. Terrain Elevation and Spatiotemporal Coherence of Runoff

The results of the mutual correlation of long-term Q series of the analyzed rivers of the Middle Volga region and the eastern part of the basin of the Don River for the baseline climatic periods under consideration are depicted in Figure 9. Several important facts are noteworthy. Firstly, this was a lower mean mutual correlation coefficient (r′) for the rivers in the Middle Volga region compared to the studied area of the Don River basin. Secondly, there was little or no change in this mean coefficient between the baseline periods. Thirdly, all partial coefficients of mutual correlation for the rivers in the Don River basin turned out to be comparatively high and statistically significant.

It is also necessary to note the presence of two “anomalous” rivers in the correlation matrix of the rivers of the Middle Volga region in the baseline period 1991–2020, the Krasnaya River (no. 7) and Bol’shoy Karaman River (no. 20) (these rivers were not distinguished by such anomalies in the previous period, 1961–1990), which were distinguished by noticeably low partial mutual correlation coefficients compared to other rivers and, in general, a mean correlation coefficient (0.49) for this group of rivers. In this regard, it is worth noting two circumstances that most likely predetermined the above-mentioned anomalies. Both the forest–steppe Krasnaya River and steppe Bol’shoy Karaman River have the lowest average basin elevations (and most likely the lowest average slope gradients in their basins) in their landscape groups (see Table 1), located within the deposits of the left-bank complex of alluvial terraces of the ancient Volga River valley, and are composed primarily of alluvial and other loams on the surface.

It is noteworthy that, according to the values of the partial coefficient of mutual correlation, the most representative rivers in both baseline periods were the Bol’shoy Cheremshan (Middle Volga region) and Khopyor (Don River basin) (see Figure 9). Above, in Figure 3, we presented graphs of changes in the flow rate of these rivers for the periods under consideration. It is also worth paying special attention to the fact that in the Middle Volga region, the basin of the representative Bol’shoy Cheremshan River is located next to the basin of the “anomalous” Krasnaya River (see Figure 1), which indicates the intra-basin causes of this anomaly (or representativeness).

The matrices of mutual correlations of long-term Q_max series of the analyzed rivers of the Middle Volga region are presented in Figure 10. Notably, the mutual correlation coefficient for Q_max increased during 1991–2020 compared to 1961–1990 (see Figure 10), indicating a rise in the spatiotemporal coherence of this runoff in the study’s subregion.

Along with this increase, there was a decrease in the number of statistically insignificant and low partial correlation coefficients. Being anomalous with respect to the annual average runoff in 1991–2020, the Krasnaya and Bol’shoy Karaman rivers did not show obvious anomalous properties with respect to the temporal variability of Q_max during the same period, although the average coefficients of their mutual correlation with other rivers were noticeably lower than the mean coefficient (r′) in this subregion (see Figure 10). As for the rivers in the eastern part of the Don River basin, the spatiotemporal coherence of the annual maximum runoff between the baseline periods remained almost unchanged (according to r′), being considerably higher and fully statistically significant than in the Middle Volga region (see Figure 10).

Similar mutual correlation procedures were also applied to long-term series of minimum runoff in the cold season (Q_min-CP) of the year, and their results are depicted in Figure A3 in Appendix A. For the rivers in the Middle Volga region in both baseline periods, a relatively low spatiotemporal coherence of their minimum runoff in the cold season of the year was characteristic (0.24–0.26, according to the mean mutual correlation coefficient) (see Figure A3). Moreover, the percentage of statistically significant partial correlation coefficients was very low (1/3 or less of all cases). In 1961–1990, rivers with a near-zero or negative average correlation coefficient were even identified (the forest–steppe Tushonka River (no. 9) and steppe Chapayevka River (no. 15)). During the period 1991–2020, the previously mentioned Krasnaya and Bol’shoy Karaman rivers, along with some other rivers, also remained relatively anomalous (r ≤ 0.1) in relation to this runoff, although in the previous period, 1961–1990, they were more representative.

The spatiotemporal coherence of minimum runoff during the cold season in the eastern Don River basin remained higher than in the Middle Volga basin, but experienced a decreasing trend from 1961–1990 to 1991–2020 (see Figure A3): from 0.50 to 0.37, according to the coefficient r′. In the same temporal direction, the percentage of statistically significant partial correlation coefficients in the specified correlation matrix decreased.

The results of the mutual correlation analysis of the series of minimum runoff during the warm season (Q_min-WP) of the year are presented in Figure A4 in Appendix A. As in the case of Q_min-CP, the spatiotemporal coherence of Q_min-WP was relatively low, especially in the Middle Volga region. For Q_min-WP, it even tended to decrease with a decrease in the percentage of statistically significant partial correlation coefficients. Again, as in all previous cases, spatiotemporal Q_min-WP coherence was higher in the analyzed Don River basin’ rivers, with a higher percentage of statistically significant partial mutual correlation coefficients, which tended to increase from 1961–1990 to 1991–2020 (see Figure A4).

To summarize the above, two principal intermediate conclusions can be made:

• During both studied baseline climatic periods, the spatiotemporal coherence of all the examined runoff parameters was higher in the rivers of the Don River basin than in the rivers of the Middle Volga region. These differences were statistically significant, except for Q_min-WP variability in 1961–1990 (Figure 11).
• For all the runoff parameters examined, the differences in subregion mean mutual correlation coefficients between 1961–1990 and 1991–2020 were statistically insignificant, except for Q_max variability in the Middle Volga region (see Figure 11).

In the identified interperiod differences in the spatiotemporal coherence of the examined runoff parameters of the analyzed rivers, a certain role was also played by the difference in absolute terrain elevation in their basins. Statistically significant relationships in this regard were manifested in the annual average (Q) and maximum (Q_max) runoff, and only in the Middle Volga region’s rivers, where these relationships were of a direct nature (Figure 12). There was no statistically significant effect of the relief factor H on the other runoff parameters examined.

4.3. Terrain Elevation and Intra-Annual Runoff Correlation

An analysis of the correlation of the long-term series of the considered runoff parameters with each other for the baseline periods allowed us to identify the following principal patterns:

• During both baseline periods, a high correlation between Q and Q_max was observed, especially in the rivers of the steppe zone, where its coefficients were 100% statistically significant (Figure 13). This primarily indicates the still crucial role of spring snowmelt-induced flood runoff, the extreme indicator of which is the maximum runoff, in the formation of the annual runoff of the region’s rivers. Moreover, in the forest–steppe zone, there was even a relative increase in this role in 1991–2020 compared to the previous 30-year baseline period, according to the increased average correlation coefficient.

• The influence of the low-water runoff of the warm season (the extreme indicator of which is Q_min-WP) on annual runoff formation in both landscape zones can be considered relatively high (see Figure 13). The influence of low-water runoff during the cold season (according to Q_min-CP) on the formation of Q was less. However, it should be noted that the correlation of the minimum runoff of both cold and warm seasons with Q tended to increase in the forest–steppe zone in 1991–2020, which also indicates an increment in the influence of these two runoffs on annual runoff formation in this zone during the period of progressive climate warming in recent decades.
• The correlations between Q_max and the minimum runoff in both cold and warm seasons were generally positive, but in most cases statistically insignificant and relatively small in magnitude (see Figure 13). However, in most cases, this positive correlation tended to increase from the first baseline period to the second period.
• A positive correlation between Q_min-CP and Q_min-WP was also observed, which was statistically significant in most cases in both baseline periods, particularly in the forest–steppe. However, while it strengthened from the first to the second period in the forest–steppe, it weakened in the steppe environment (see Figure 13).

In the steppe zone, in both baseline periods, direct relationships, r = f(H), were maintained in the correlated pairs Q and Q_min-CP, as well as Q and Q_min-WP, although they were statistically insignificant (Figure 14). In the forest–steppe zone, these relationships were also direct, but already statistically significant, only for the correlated pairs Q and Q_min-WP in both periods, and for the pair Q and Q_min-CP in the period 1961–1990: the higher H, the greater the contribution of low-water runoff (represented by one of its indicators—minimum runoff) to the overall variability of the annual runoff.

It is noteworthy that for the correlated pair Q and Q_min-CP, the positive relationship, r = f(H), for 1961–1990 changed to the opposite in 1990–2020 under the conditions of progressive climate warming; this relationship, however, did not have statistical significance. In other words, the role of cold season low-water runoff in the formation of annual runoff in 1990–2020 was probably somewhat greater in lowland river basins than in upland ones.

As for the relationship r = f(H) in the correlation of Q_max and Q_min-CP, the same patterns were observed there as in the correlation of the pair Q and Q_min-CP, but with the only difference that this relationship was statistically significant in 1991–2020, and insignificant in 1961–1990 (Figure 15). In other words, the relationship between long-term changes in Q_max and Q_min-CP during the period of active climate warming (1991–2020) was higher in lowland forest–steppe landscapes than in upland ones.

For the remaining pairs of correlations (Q and Q_max; Q_min-CP and Q_min-WP), no statistically significant relationships, r = f(H), were found across the landscape zones and baseline periods under consideration. The results of these correlations are depicted in Figure A5 and Figure A6 in Appendix A. However, the absence of any patterns in the correlations of these runoff parameters with the elevation of the terrain is also a very useful result, allowing us to state, among other things, the prevailing role of other unaccounted factors, primarily the so-called “active” driving forces that control the intra-annual restructuring of the runoff of the study region’s rivers.

5. Discussion

The described trends in the influence of the elevation factor on the annual average specific runoff—Z (see Figure 4)—are logically explained by the behavior of this factor’s influence on snowmelt flood runoff and low-water runoff. Thus, in the forest–steppe zone, increasing low-water runoff during the cold and warm seasons as the average elevation of river basins increases is more or less offset by a decrease in snowmelt flood (including maximum) specific runoff, resulting in weak and statistically insignificant dependences, Z = f(H). On the contrary, in the steppe zone, the well-defined direct dependence, Z = f(H), was due to the absence of such compensation from the snowmelt flood runoff, which, as is shown in Figure 5 for Q_max, was indifferent to changes in H.

What are the reasons for such a behavior of high-water specific runoff with an increase in terrain elevation in the forest–steppe zone? Since most of the analyzed rivers flow in the Middle Volga region, the identified relationships, Z_max = f(H), are largely due to the peculiarities of snowmelt flood runoff formation in this subregion. Without going into detail, it is important to note that the Middle Volga region is characterized by a well-defined tiered structure (up to 2–3 plateau-like levels—ancient relief planation surfaces) [33], which is clearly reflected in the character of the natural and natural–anthropogenic landscapes of the forest–steppe. The climate of the upper planation level (more than 200–300 m a.m.s.l.) is “northernized” compared to the climate of the lower levels (less than 100–200 m a.m.s.l.), since there is more precipitation there (by about 12%), and the annual average (by more than 1 °C), summer and winter air temperatures are lower [36]. This “northerning” of the upper plateau’s climate, coupled with the spread of siliceous, carbonate-free Paleogene rocks on the surface (for example, in the Middle Volga region’s Pre-Volga subregion), favors soil podzolization processes (various subtypes of gray forest soils). Consequently, pine, pine–broadleaf, and deciduous forests form on the upper plateau’s surface. Due to the low soil fertility, forests at these levels have not been severely cut down, but occupy a relatively high percentage of the upper plateau’s surface. A different landscape is characteristic of lower elevations. There, under relatively drier climates and more fertile soils (Chernozems and humus–carbonate soils), open, plowed areas, previously primarily steppe, predominate [34]. Thus, in this region of Eastern Europe, the structural features of natural landscapes and their anthropogenic changes are closely dependent on geological and geomorphological conditions.

The above means that in river basins located within, or mostly on, the upper plateau, in relatively cooler and wetter climates, especially those with abundant forest vegetation, spring snowmelt is extended over time (especially on northern-facing slopes) compared to lower-lying river basins. Consequently, snowmelt floods there are also extended, and flood peaks are lower (lower Q_max) (see Figure 5). Furthermore, forest vegetation facilitates the extending of snowmelt time and the transformation of melted surface runoff into groundwater. Conversely, higher flood peaks in lower-lying river basins (see Figure 5) are facilitated by considerable plowing and deforestation there. Apparently, these trends are maintained during the transition from spatial variability of Q_max to its temporal variability: between 1961–1990 and 1991–2020, the largest losses in Q_max occurred in lower-lying river basins (see Figure 8), where regional climate warming appears to have been greater than at higher elevations. Despite the average area of the analyzed forest–steppe river basins in the “high-elevation group” (>153 m a.m.s.l.) was more than twice as large as that of the basins in the “low-elevation group” (≤153 m a.m.s.l.) (see Table 1), this did not in any way affect the identified dependencies for Z_max = f(H), since Q_max was determined in specific units of runoff.

The trends in the reduction in maximum runoff Q_max between the studied climatic periods shown in Figure 8 fit well into the general context of the current reduction in snowmelt-induced runoff revealed by the results of stationary observations in the study region. Thus, as shown by the results of long-term observations at the experimental Novosil zonal agroforestry melioration station, located in the southwestern sector of European Russia (the forest–steppe zone of the Central Russian Upland), for 1959–1974, snowmelt-induced runoff on the station’s slopes was absent for only one year, with a maximum in other years of 146–186 mm depending on the nature of cropland cultivation. During 1975–1997, snowmelt runoff was absent for seven years, with a maximum of 40–50 mm. In the period 1998–2022, there was almost no snowmelt runoff (with a maximum of 46–71 mm) [37]. The lack of meltwater runoff on the slopes was due to weak or no soil freezing, as well as low soil moisture immediately before the formation of snow cover in the late fall season. If the soil has thawed or frozen slightly (up to 50 cm), meltwater runoff does not occur. In conditions when snow falls on deeply frozen (>50 cm) and very moist (>120 mm) soil in the late fall season, meltwater runoff always occurs in the spring and is determined by snow reserves and soil moisture [37]. Taking into account the altitudinal zoning of the climate in the region, which is particularly well expressed as the continentality of the climate increases to the east, and the fact that the results obtained at the Novosil station characterize the upland landscapes of the East European Plain (about and more than 200 m a.m.s.l.), it can be assumed that in the conditions of the low-lying (relatively warmer) landscapes of the study region, the degradation processes of melt runoff are even more catastrophic in the context of progressive climate change. All these issues require comprehensive study in representative river basins.

As for the steppe zone, the overall zero response of Q_max to changes in H (see Figure 5) can be partially explained by the less explicit “northernization” of landscapes, or its absence, which is expressed in much greater plowing of uplands and a significantly smaller forest area than in the forest–steppe zone. A certain reduction in landscape–climatic differences in the elevation of river basins in this zone is also facilitated by an overall decrease in the average elevation of the considered uplands towards the Caspian Lowland [33,34] (see Table 1). These assumptions require confirmation by further research in individual catchments in conjunction with field work.

The direct and close relationship between H and Q_min-CP is also explained by the “northernization” of the climate and landscapes of uplands (see Figure 6). Higher winter precipitation and, consequently, deeper snow cover (which is also facilitated by forested areas, especially in forest–steppe zones) at high elevations facilitate greater input of melted snow into the underground component (groundwater runoff). In recent decades, an increase in the number of winter thaws has been observed in the study region, caused by an increment in the sum of positive air temperatures in the cold season of the year [38]. This circumstance further strengthened the relationship Q_min-CP = f(H) during the period of observed progressive global warming, 1991–2020, compared to 1961–1990 (see Figure 6). In part, the strengthening of this relationship could also be due to a slight increment in winter precipitation both in the Don River basin [39] and Middle Volga region [40]. It is for the reasons listed above that an increase in minimum runoff of the cold season in 1991–2020 compared to 1961–2020 was observed in the analyzed upland rivers, primarily in the steppe zone (see Figure 9). We also cannot rule out the possibility that upland rivers, with their more deeply incised valleys, may be fed by water from a larger number of underground aquifers, further increasing their water runoff in the low-water seasons. However, to resolve this issue, a thorough study of the geological and hydrogeological structure of the valleys themselves is necessary in each specific case.

The same altitudinal landscape–climatic differentiation underlies the direct dependence, Q_min-WP = f(H), with the only amendments being that the river’s supply in the summer–fall low-water period (warm season) is associated with replenished reserves of groundwater both during the period of active spring snowmelt and summer–fall rains. This river feeding pattern during the summer–fall low-water season most likely has a complex nature of a combination with the relief and geological (and hydrogeological) structure, as a result of which the relationship between changes in H and Q_min-WP is generally less evident and statistically insignificant compared to Q_min-CP. It is worth noting again the strengthening of this relationship in 1991–2020 (see Figure 7) in the steppe zone’s rivers, where there has been an increase in the number of heavy rains in recent decades [22], which apparently make their own certain contribution to the increase in Q_min-WP with terrain elevation.

Geomorphologically tiered and associated landscape–climatic differentiation are most likely the main reasons for the increased spatiotemporal coherence of the annual average and maximum runoff in the Middle Volga region with increasing elevation compared to the eastern part of the basin of the Don River. This is due to similar patterns of response of natural and natural–anthropogenic landscapes of different altitude zones of the study region (primarily the Middle Volga region) to modern climate change, primarily the increase in winter air temperatures and the frequency of winter thaws, as well as precipitation and evaporation, their ratio, etc. [38,40]. These patterns are also influenced by lithologic features and land use changes in recent decades in each individual basin, making them difficult to interpret at this study’s stage. In general, the most coherent rivers in 1991–2020 with respect to the annual average runoff were those that had the highest long-term mean values of specific runoff (on average 3.8 L s^–1 km^–2, upper-lying river basins with H > 132 m a.m.s.l.), while the least coherent were rivers flowing through lowland landscapes (on average 1.9 L s^–1 km^–2, lower-lying river basins with H ≤ 132 m a.m.s.l.) (see Figure 12 and Table 1).

Summarizing the above with respect to the analyzed landscape zones, it can be stated that the most striking differences in the influence of relief on specific runoff trends in the region’s rivers were manifested between the forest–steppe and the steppe in annual average and annual maximum runoff. We assume that this associated, first of all, with a more distinct reaction of the runoff of forest–steppe rivers to a more noticeable climate change in the cold season within more northern (forest–steppe) landscapes than in more southern (steppe) landscapes in the region [40], of course, taking into account the previously mentioned differences in the landscapes of these zones (including their “northernization”).

6. Conclusions

The achieved results highlight the high importance of relief, especially its altitude characteristics, as a driver of the spatiotemporal variability of runoff in small and medium-sized river basins at the regional level. The driving role of the terrain elevation factor was distinctively manifested in the changes in the annual average runoff in the study region’s steppe landscape zone, and in the annual minimum runoff of the cold season of the year in both baseline periods considered. The influence of this factor on the spatiotemporal coherence of the runoff (the annual average and maximum runoff) of the analyzed rivers in the key subregion of the study region—the Middle Volga region—was also noticeable. At the same time, the results obtained from the Don River basin should be treated with caution due to the comparatively small number of rivers analyzed in it.

It is very important to understand that the role of the elevation factor, especially its average value across river basins (H), does not manifest itself in isolation from other factors. On the contrary, the average elevation of river basins is a key morphometric characteristic, serving as an integral indicator that sets the “energy” and “moisture” background for landscape formation. Changes in the H basin entail restructuring of the entire natural complex, even within relatively low plains. It is through its landscape-regulating role that terrain elevation controls the spatial variability of river runoff, as well as its temporal transformations, including in the modern era of active climate warming. In addition, relief characteristics are closely integrated with other “passive” driving forces that control river runoff, primarily the lithologic and tectonic structure, which can, to one degree or another, transform the regional patterns of its spatiotemporal changes.

As was shown earlier [20,21,22,29,30,31,32,41,42], progressive climate change in recent decades has led to extensive regional changes in river runoff, which were expressed primarily in the transformation of its seasonal components, which resulted in a decrease in intra-annual runoff changeability, primarily in the East European Plain’s southern half and the plain’s western part (including the Baltic States, Belarus, etc. [31,43]). Similar trends can also be seen towards the east, in the northern part of Kazakhstan [44]. All these changes were reflected, among other effects, in the same extensive reductions in soil and gully erosion intensity and associated river sediments in the region [27,32,45,46,47,48]. Under these conditions, taking into account all factors, and not just “active” ones, such as climate and land use, will provide a reliable basis for high-quality forecasting of future changes in river runoff and their socio-environmental consequences. The results presented in this article are intended to make a feasible contribution to the understanding of the factor functions of relief in the (trans-)formation of modern and future runoff in river basins.