Interannual variations in annual air temperature (Ta) in the NE-Amur, the NW-Amur, and the S-Amur between 1960 and 2000 exhibit similarities (Fig. 2a). No abnormal Ta was found in the three regions between 1995 and 1997, when the extreme increase in dFe concentration in the Amur River was recorded. On the whole period, peaks in Ta seemed to occur once every few years; they occurred in 1963, 1968, and 1975. In particular, high Ta was maintained for three years between 1988 and 1990 throughout the entire Amur River basin, with 1990 being the warmest year between 1960 and 2000 in the NE-Amur and S-Amur. Interannual variations in accumulated temperature (AT) in the three regions also exhibit the similar variation to those of Ta (Supplementary Fig. S2). During the warm years between 1988 and 1990 in the Amur River basin, the highest and the second highest AT was also found in 1988 and 1990, respectively, in the NE-Russia.
Late summer net precipitation (P − E) between 1960 and 2000 in the NE-Amur, the NW-Amur, and the S-Amur also exhibit similar interannual variations (Fig. 2b). It is interesting to note that late summer P − E in the Amur River basin during this period was clearly seemed like interdecadal variation. P − E in the three regions decreased after 1960; after around 1977, P − E increased and remained higher positive values for the following 20 years; increased P − E suggests that the soil in the Amur River basin has become wetter since 1977 (and since 1980 for the S-Amur). Since 1977, P − E in the NE-Amur and the NW-Amur has been higher than that in the S-Amur; this suggests that there have been larger increases in soil moisture in the northern parts of the Amur River basin than in the southern regions.
For the variations in late summer P − E between 1960 and 2000 in the NE-Amur and the NW-Amur, significant correlation was found with 7-year moving average of the PDO index (NE-Amur r2 = 0.27, p = 0.0014; NW-Amur r2 = 0.77, p = 6.35 × 10− 12) (Fig. 3a). This indicates that the increased P − E after about 1977 was related to the change in PDO phase from negative to positive in 1977. As a result of examining the anomaly of geopotential height at 500 hPa during the positive phase of the PDO (1977 − 1997), there was an anticyclonic anomaly around the Sea of Okhotsk and the Bering Sea, and a cyclonic anomaly around the NW-Amur (Fig. 3b). These results are in close agreement with those from Zhang et al. (2021)43, which showed the presence of anticyclonic and cyclonic teleconnection wave lines at the midlatitudes of the Northern Hemisphere during the positive PDO phase. Such changes in the atmospheric circulation could have resulted in increased water vapour flux convergence into the upper part of the Amur River basin in late summer after 1977.
Figure 4 shows the interannual variations in annual/seasonal dFe concentrations in the Amur River. There were extreme increases in annual and late summer dFe concentrations between 1995 and 1997 (Fig. 4a,c), which occurred 1 year earlier than the large increases in late spring dFe concentrations between 1997 and 1998 (Fig. 4b). Cross correlation coefficients between dFe concentration and climate variables are shown in Fig. 5. There was significant positive correlation between annual dFe concentrations and Ta with lags of 7 (all regions: r = 0.43–0.55, p < 0.01) and 8 (NE-Amur: r = 0.48, p = 0.0021; S-Amur: r = 0.44, p = 0.0052) years (Fig. 5a). Comparing the interannual variations in annual dFe concentrations and Ta, it certainly looks like that the peaks in annual dFe concentrations in 1970, 1975, 1982, and 1995–1997 (Fig. 4a) occurred 7 years after the peaks in Ta in 1963, 1968, 1975, and 1988–1990 (Fig. 2a). There was also significant positive correlation between late summer dFe concentrations and Ta with a lag of 7 years (all regions: r = 0.54–0.69, p < 0.01), and the correlation coefficients were larger than those between annual dFe concentrations and Ta (Fig. 5c). In addition, there was significant positive correlation between late spring dFe concentrations and Ta with lags of 8 (all regions: r = 0.44–0.47, p < 0.01) and 9 (all regions: r = 0.47–0.57, p < 0.01) years (Fig. 5b). It is also worth noting that the correlation coefficients between annual/seasonal dFe concentrations and Ta were the highest in the NE-Amur; this indicates close association between high Ta in the NE-Amur and increase in annual/seasonal dFe concentrations 7 years later (9 years for late spring) in the Amur River. The similar relationship was also found between annual/seasonal dFe concentrations and AT in the NE-Amur with a lag of 7–9 years (Supplementary Fig. S3). In particular, the highest correlation coefficient was found between late summer dFe concentrations and AT in the NE-Amur with a lag of 7 years (r = 0.51, p = 0.0010) (Supplementary Fig. S3c). Considering that AT is commonly referred to as the index of maximum active layer thickness44, this lag relationship indicates that increase in late summer dFe concentrations in the Amur River was closely associated with permafrost degradation under a warm summer 7 years ago in the NE-Amur. In contrast to the results between dFe concentration and Ta, there was no significant correlation between dFe concentrations and P − E (Fig. 5d–f).
During a warm year, permafrost degradation results in thawed soil, and the Fe(III) minerals in the thawed soil can be exposed to microbial reduction, which could lead to the intensive generation of soluble Fe(II) in the deep soils in the active layer. According to Herndon et al. (2015)20 which investigated the vertical distribution of dissolved Fe(II) in the active layer in summer, dissolved Fe(II) concentration was the highest in the deep soils near the permafrost table. Abundant Fe(II) in deep soils can potentially increase riverine dFe concentration by forming complexes with humic substances, which are soluble in river waters with neutral pH9. However, because soil permeability is low, the Fe(II) in deep soils will take several years to travel through the deeper part of active layer to the rivers21,45,46. This lag between the intensive generation of Fe(II) and the appearance of the iron in the rivers is supported by the significant correlation between annual dFe concentration and Ta with a lag of 7 years (Fig. 5a) and that between late summer dFe concentration and Ta (Fig. 5c and Fig. 6a). In other words, the strong correlation between annual/late summer dFe concentration and Ta in the NE-Amur with a lag of 7 years indicates that intensively generated Fe(II) in a warm year discharged to river 7 years later through the deeper soils in the active layer. In particular, the strong correlation between late summer dFe concentration and Ta can be reasonable for this mechanism because the active layer thickness is at its maximum and permafrost-affected rivers are mostly dominated by deep groundwater47,48 in late summer. Given the high correlation coefficients between annual/late summer dFe concentration and Ta with a lag of 8 years, we infer that increased dFe discharge likely lasted for 1–2 years. In addition to input of groundwater rich in dFe, summer rainfall is also considered to increase dFe concentration in rivers17,49. However, there was no significant correlation between the individual measurements of the Amur River discharge and dFe concentration in late summer (Fig. 6b). Unexpectedly, significant correlation was found between late spring dFe concentration and Ta with lags of 8 and 9 years (Fig. 5b and Fig. 6c) even though the influence of groundwater is minimum during spring floods. One possible reason is that precipitated Fe(II) as iron-oxyhydroxides after flowing into the river dissolves again the following spring when it is in contact with the large amounts of DOC released from the organic-rich topsoil by snowmelt. Another possible reason is that deep groundwater is not fully frozen and continues to interact with river waters even in spring. Although it is well known that dFe concentration increases when spring floods peak50,51, river discharge is inadequate to explain late spring dFe concentrations in the Amur River (Fig. 6d).
On the basis of our results, we hypothesize that the extreme increase in dFe concentration in the Amur River between 1995 and 1997 was directly caused by the continuous warming between 1988 and 1990 in the Amur River basin (Fig. 2a), especially warm summer in the NE-Amur in 1988 and 1990 (Supplementary Fig. S2). The continuous warm years resulted in permafrost degradation, as an increases in soil temperature were observed on 10 weather stations in the NE-Amur between 1988 and 199018. Permafrost degradation and subsequent generation of Fe(II) in the deeper part of the active layer could have promoted iron bioavailability, and led to the extreme increase in riverine dFe concentration 7 years later (1995–1997). In addition to the warm years between 1988 and 1990, increased late summer P − E after 1977 (Fig. 2b), resulted from increased water vapour flux convergence during positive PDO phase (1977–1997), could also have promoted dFe discharge to the Amur River between 1995 and 1997. Although there was no significant relationship between the interannual variations in annual/seasonal dFe concentrations and late summer P − E, persistent positive P − E in late summer after 1977 in the Amur River basin may include the following effects on biogeochemical and hydrological cycles: (1) intensification of permafrost thaw and degradation because of increased water infiltration and larger heat transport to deeper soil52, (2) promotion of microbial iron reduction under anaerobic conditions because of increased soil water content53, and (3) increased groundwater discharge to rivers54. These effects were likely more intense in the NE-Amur and the NW-Amur, which showed higher positive values of late summer P − E after 1977. On the source region of dFe to the Amur River, the previous studies provided the important information by investigating the distribution of riverine dFe concentration in the middle and lower Amur River basin; the amount of dFe, humic acids, and organically-bound Fe was much higher in rivers in the NE-Amur than rivers in the S-Amur because taiga forests and lowland wetlands in the NE-Amur have a large amount of organic matter and humic substances55,56. Unfortunately, there is no information on riverine dFe concentration in the NW-Amur, but it is supposed to be low due to low concentrations of DOC and humic substances in rivers in this region57. Taking into account these chemical characteristics of river water in each region and the strongest correlation between annual/seasonal dFe concentrations and Ta in the NE-Amur (Fig. 5a–c), permafrost-covered area of the NE-Amur is probably the most important region to occur the extreme increase in dFe concentration between 1995 and 1997.
The results of our study highlight the importance of the interactions between high Ta and late summer P − E in iron biogeochemistry; these interactions should be taken into consideration in the study of the biogeochemistry of iron in the Amur River basin. However, there is still a huge lack of knowledge about dFe discharge mechanism through the deeper part of the active layer. To better understanding the extreme increase in dFe concentration in the Amur River between 1995 and 1997, field studies in the NE-Amur are needed to examine the permafrost distribution, soil profiles (peat and mineral horizons), soil permeability, and dFe movement in the deeper part of the active layer. Moreover, numerical modelling studies are also needed to validate the time lag relationship between the annual/seasonal dFe concentrations and Ta that was reported in this study.