**4. Discussion**

#### *4.1. Litter Dynamics in Different Years*

Litter volume is a component of the forest ecosystem biomass, which reflects the primary productivity level of the forest ecosystem and reflects the functions of the forest ecosystem [26,67–70]. The global annual change range of forest litter is 0.13 (Alaska's Picea crassifolia forest in the United States)—15.3 (Congo's tropical rain forest) t·hm−<sup>2</sup> [28,42], and it can be divided into tropical, subtropical, temperate, and cold temperate zones according to climatic conditions. The forest litter yield in different climatic zones shows a downward trend with the increase in latitude. On average, Eurasian forests are tropical (8.40 <sup>t</sup>·hm−2) > subtropical (5.25 <sup>t</sup>·hm−2) > temperate (3.81 <sup>t</sup>·hm−2) > cold temperate (t·hm−2) [21,37,43,71]. The annual litter amount of different forest types in different climate zones in the world also varies greatly. The annual litter amount of main forest types in different climate zones can be shown as the maximum average annual litter amount of tropical rain forest and seasonal rain forest, which is 9.98 t·hm−2. It is followed by evergreen broad-leaved forest, with an average annual litter of 6.96t·hm−2. Then, 5.79 t·hm−<sup>2</sup> in coniferous and broad-leaved mixed forest; deciduous broad-leaved mixed forest 5.1 <sup>t</sup>·hm−2; The coniferous forest with the smallest litter is 4.77 t·hm−<sup>2</sup> [72]. The subtropical evergreen broad-leaved forest in the study area is located in the north-south transition zone between the middle subtropical climate and the south subtropical climate. In this study, the average annual total litter in 2005, 2010, and 2015 is 8.18 <sup>t</sup>·hm−2, which is within the range of subtropical forest changes (1.01–13.00 <sup>t</sup>·hm−2). Compared with the evergreen broad-leaved forest in different regions of the subtropical zone, its annual litter is less than that of the Dinghushan monsoon evergreen broad-leaved forest in the south subtropical zone (8.45 <sup>t</sup>·hm−2) [17], greater than that of the Tiantong evergreen broad-leaved forest in the north subtropical zone (5.55 <sup>t</sup>·hm−2) [73–75], but similar to that of the Xiaokeng subtropical evergreen broad-leaved forest in the middle subtropical zone (7.99–8.450 <sup>t</sup>·hm−2), It shows that the litter of wet evergreen broad-leaved forest in Ailao Mountain is similar to the geographical location and stand composition structure of the middle subtropical forest [70].

In different years, each component of litter and its proportion in the total forest litter are the largest, and the sum of the two reaches 76.17% of the total, which is consistent with the general situation of plant growth and litter, and basically consistent with the proportion sequence of each component of litter in other research results [71,76–79]. The total amount of litter in the study area in 2005 (normal year without extreme weather interference), 2010 (affected by extreme arid weather), and 2015 (affected by extreme ice and snow weather) are 8.82, 7.70, and 8.01 <sup>t</sup>·hm−2, respectively, of which the proportion of litter is 49.96%, 61.52%, and 41.30%, respectively, and the proportion of litter branches is 25.82%, 17.26%, and 32.05%. Compared with other subtropical forests of the same type, the annual average percentage of leaves (50.89%) is lower, while the three-year average percentage of branches (25.28%) is higher. In 2005, the proportion of litter branches in the study area was less than 30%, which was still in the middle and early stages of succession. The proportion of litter branches was small. With the succession, the proportion of litter branches should have gradually increased, but the proportion of litter branches in the region from 2005 to 2010 was gradually decreasing, reached the valley in 2010, and then gradually increased from 2010 to 2015. This is because of the rare serious drought in Yunnan in 2010. After the forest was disturbed by the extreme arid weather for a long time, the characteristics of the plant community degenerated toward the type of low-grade community, and secondary succession occurred within the community. At the same time, because the soil was lack of water for plant growth in this dry season, some leaves fell, so the proportion of branches and litter was less than that when the forest was not disturbed, The proportion of leaf litter also increased correspondingly [42,43], indicating that extreme drought weather interference seriously hindered the normal succession process of forests. In 2015, the proportion of litter leaves was less than that of undisturbed years, but the proportion of litter branches increased accordingly, which is consistent with the short duration of this extreme ice and snow weather disturbance. The wood litter in the forest, such as branches, is quite random. The litter collected every month is the litter branches that died on the trees before. Therefore, the litter of branches is usually not directly related to phenology but is greatly affected by climate factors. It also shows that although the evergreen broad-leaved forest in Ailao Mountain is affected by extreme ice and snow weather, However, due to the rich species and complex community structure in Ailao Mountain, the forest ecosystem has the ability of self-regulation and high early post-disaster recovery, which makes the evergreen broad-leaved forest vegetation in Ailao Mountain recover slowly. Thus, the litter amount of evergreen broad-leaved forest in Ailao Mountain after being disturbed by extreme weather is close to that of evergreen broad-leaved forest in the same latitude in normal years [37,71]. To sum up, the impact of extreme drought weather on litter yield and nutrients in the study area is greater than that of extreme ice and snow weather interference, and it mainly has a greater impact on litter branches, litter leaves, and the total amount of litter, while the impact on other litter components is relatively small [56].

Due to the influence of external environmental factors, the dynamic changes of litter in different years show some regularity but also show some differences [70–72]. The cause of the occurrence of the litter rhythm is mainly determined by the climate change factors and the biological characteristics factors of the tree species that make up the community [1,3,23,47]. The monthly dynamic performance patterns of litter in 2005, 2010, and 2015 are different, including single peak, two peaks, and four peaks, with different peak sizes, and all of them play a leading role in litter amount. In the normal year 2005, the maximum peak value of withering occurred in the rainy season, while in the two special years, 2010 and 2015, the peak value of withering mainly occurred in the dry season.

In 2005, the total amount and each component of litter reached the maximum in August, which was affected by the external temperature and accelerated the rate of metabolism, prompting a large number of plants to litter. The temperature in August is suitable for the vigorous growth of plants. The rising temperature of evergreen trees during the germination period will promote the sprouting of new leaves and accelerate the aging and apoptosis of old leaves. When the nutrients required for the growth of new leaves exceed

the nutrients absorbed from the soil, the plants will preferentially transfer the nutrients stored in the old leaves for the growth of new branches and leaves, and the transfer and reabsorption of nutrients will accelerate the apoptosis of old leaves [17,39,71]. The other peak months in 2005 are basically the same as those in 2010. Although the plants in the evergreen broad-leaved forest gradually change leaves throughout the year, the temperature rises at the beginning of the rainy season (April and May), and a large number of new leaves germinate and grow vigorously. In these growing seasons, new leaves compete with old leaves for limited mineral elements. Because the strong vitality of new leaves makes the elements that can be transported from old leaves to new leaves, and finally causes the senescent leaves to fall off one after another, During this period, the rainfall increased, so the first short and concentrated peak of leaf litter appeared in April, basically in line with the physiological characteristics of evergreen broad-leaved forest. Another secondary peak of litter in the study area is at the end of the rainy season in November. The occurrence of the second peak of litter is related to the defoliation period of the dominant tree species of the evergreen broad-leaved forest in this area, Kaempferous, and because of the sudden drop in humidity and temperature in autumn and winter, the leaves of some evergreen broad-leaved forests lose their vitality in dry air and wither [17,70,71].

In January 2015, the study area had an extremely strong ice and snow weather. Because a large number of leaves on the crown of evergreen tree species condensed with ice and snow when the ice and snow disaster occurred, and some tree species were easy to break, strong wind, heavy rain and heavy snow can blow and break the litter branches of the previous period to the ground [5,14,37,42,70], After the interference of extreme ice and snow weather, the leaves were violently shaken by strong external forces, resulting in non physiological shedding, which led to a sharp increase in the amount of litter in January 2015 and a peak. This randomness can cause grea<sup>t</sup> changes in the number of litter branches at different times, which also indicates that the weather conditions in 2015 were worse than those in 2005 and 2010. In addition, after the interference of extreme ice and snow weather, a large number of trees fell down, branches were crushed and dropped by heavy snow, forest canopy was seriously damaged, forest canopy density was reduced, and insect food in habitats with severely damaged vegetation was reduced, so the amount of litter brought by insects was reduced successively, thus significantly reducing the amount of litter after the peak in January. However, the rose fluctuation amplitude of the litter after the forest was disturbed by the extreme ice and snow weather decreased significantly and tended to be stable, indicating that the extreme ice and snow weather caused grea<sup>t</sup> damage to the physiological structure of the tree species, leading to changes in the subsequent litter patterns. In a word, the litters in the three years have their own distinct littering rhythms, which may be because the species composition of evergreen broad-leaved forests in the study area, the climatic conditions and the degree of interference vary greatly in different years, and the litters are very vulnerable to the biological characteristics of forest species, climatic conditions, and other environmental factors, making the seasonal dynamics of each year different [20].

#### *4.2. Nutrient Concentration Dynamics and Nutrient Return of Litter*

The return of nutrients from the litter to the soil maintains soil fertility and promotes nutrient cycling in forest ecosystems, which is important for improving the habitat conditions of forest trees [2,26–30]. Analyzing the dynamics of nutrient concentrations and return in the litter is necessary for understanding the function of forest ecosystems [5,8,9,31,43,74,80].

The nutrient content of litter can reflect the nutrient utilization efficiency of plants. In this study, the total amount of litter and the different nutrient concentrations of each component have the same order, with the highest content of C, followed by the content of Ca and N, and the content of K, Mg, S, P, and other elements is relatively low [14,39,48]. The formation of this sequence is because the fresh litters in the study area are in the early stage of decomposition, and the initial content of N element is high, while the elements

with low content of P and K in the litters are reusable elements, which can be transferred in large quantities before the leaves fall, and then the nutrient reuse is realized through the plant transfer mechanism, which indicates that the evergreen broad-leaved forest plants in this area have high nutrient utilization efficiency. However, the strong seasonal dynamics of nutrient concentration of each component also show that the transfer amount of different elements is different in different seasons, and the decomposition rate of litter is faster due to the rapid leaching and degradation of P and K [70,71,81]. The content of nutrient elements is greatly affected by their physiological functions. The subtropical climate is hot and rainy, so the leaching effect of litter is very strong. The rainy season in this region is long, mainly concentrated in summer and autumn, and the physiological functions of each element are different [66]. Among them, K and P are easy to be lost due to rainfall, but P is a limiting factor for plant growth and development in tropical and subtropical regions. For this reason, most organisms in the subtropical ecosystem have a mechanism to maintain that P is not leached out under high temperature and humidity conditions, which makes the K and P contents in the litter lower in spring and summer as a whole, but P content is far lower than K content [3,4,26]. Ca, N, and other elements are relatively stable and not easy to be washed away by rain, which makes the seasonal variation of various nutrient elements in different years and different components inconsistent, and there is no obvious correlation. These nutrient elements will eventually return to the soil, which is of grea<sup>t</sup> significance for maintaining the long-term productivity of the forest land [82].

The chemical properties of litter are the internal factors affecting litter decomposition. C/N and C/P in the litter are common indicators of litter decomposition. N/P also has a certain characterization ability for litter decomposition. Since N can affect the growth and turnover of soil microbial communities, the C/N ratio can best reflect the rate of litter decomposition [42]. N and P control the important process of litter decomposition through coupling with C. The C/N ratio of the litter in the study area in 2005 (the normal year without extreme weather interference) is 48.65, which is greater than 25, indicating that N is the main element limiting litter decomposition in the region [51]. From 2005 to 2015, the average C/N ratio was the highest in 2005, and the content of N was the lowest, indicating that the decomposition rate of litter was the slowest in that year, and the extreme weather interference did not weaken the decomposition rate of litter. The C/N ratio of litter branches and litter leaves is not the same each year, so it can be concluded that the response of nutrient content among various organs of forest litter to extreme weather disturbance is also different [82].

The annual total amount of litter in each year is in the order of 2005 (normal year not affected by extreme weather) > 2015 (affected by extreme ice and snow weather) > 2010 (affected by extreme drought weather). From the perspective of the annual amount of nutrient return of litter, the annual amount of nutrient return is roughly proportional to the amount of litter. In addition, leaf litter is the main component of the litter, accounting for more than 50% of the total litter. Compared with other components, the nutrient concentration is generally high and easy to decompose, so leaf litter is the main body of nutrient return of the litter.
