4.1. Vegetation Change Digital Model Evaluation
Remote sensing technology provides an accurate, rapid, and cost-effective method for wetland change detection [
11]. Various digital methods can be used to detect changes by comparing two or more images of a study site from different periods, such as pixel-to-pixel comparison before and after vegetation change [
26]. The classification change detection method is one of the most appropriate and commonly used change detection techniques. This technique can easily provide a change matrix from which transfers between one vegetation type and another can be visualized. This method identifies areas of change as pixel-by-pixel differences between categories after obtaining a classification image [
29]. This can clearly show areas of change as well as transformation categories. The model proposed in this study estimates the process of vegetation change by studying different periods.
In this long-term study, based on Landsat data of the summer growth season in 2019, we detected change at three time points in 1987, 2001, and 2010. We generated a digital spatial model of vegetation change through Matlab programming and used it to detect changes in an NDVI map. The model proposed in this study has certain advantages in quantifying vegetation change in the floodplain wetlands. This evaluation method based on remote sensing imagery and Matlab modeling has good operability and detection accuracy. It is not only suitable for research on vegetation diversity in submerged areas but also provides a valuable reference for river management decision-makers pursuing wetland protection and other aspects.
4.2. Reasons for the Differences in the Spatial Distribution of Vegetation
The degree of connection between the main channel of a river and a riverine wetland will affect riverine wetland vegetation [
30]. It was found that, in the natural state, the composition and structure of plant communities, and their changes, are closely related to flooding occurrence [
31].
The overall high NDVI values observed near riverbanks indicate that most of the vegetation near these locations benefits from water recharge from the Hailar River. The fertile silt left after periodic flooding also improves the nutrient content of the soil on both banks, and water vapor improves the surrounding microclimate, thus forming a high-biomass plant belt along the riverbanks. Even plants of sandy open sand habitat with sparse vegetation (
Caragana microphylla Lam,
Cleistogenes squarrosa (Trin.) Keng, etc.) form high biomass and high canopy closure plant communities near riverbanks, which greatly weaken the mobility of sand dunes. In the horizontal direction, the vegetation type varied with distance from the riverbank. In
Figure 6, the riverbank vegetation types, in the direction from the river to the habitats more distant from the river, are forest (
Populus L.,
Salix Rosmarinifolia L. var
brachypoda (Traktv. Et Mey) Y. L. Chou,
Salix gordejevii Y. L. Chang et Skv.,
Caragana sinica (Buc’hoz) Rehder,
Pinus sylvestris var. mongolica Litv., mixed forests, etc.), grassland mixed with forest, and grassland. As shown in
Table 4, the NDVI thresholds for grassland are lower than those of forest, so the NDVI tends to decrease with distance from the riverbanks. Due to the special environment of riverine wetland systems, wetland vegetation development is associated with the river. The results show that the direct impact of the river on the vegetation on both sides of the river decreases with distance. If river runoff decreases and flood cycles become longer, it will adversely affect the wetland forest and riverine meadow ecosystems that depend on periodic flooding, leading to degradation of wetland forest areas, reductions in near-riparian wetland grassland areas, and degradation of wetland grassland systems [
32,
33].
4.3. Vegetation Change Process and Its Influences
Figure 7 shows the changes in NDVI categories occurring in the three time periods 1987–2019, 2001–2019, and 2010–2019.
Figure 7A shows that of the first three transformation categories, the changes occurring between 2010–2019 are obvious, while the conversion of forest to open sand habitat with sparse vegetation and grassland changed significantly between 1987–2019.
Figure 7B shows that within the three time periods, among the six transformation categories, the conversion of grassland to forest changed most significantly, followed by conversion of forest to grassland, indicating that more pronounced vegetation changes occurred within the three time periods.
Vegetation change in floodplain wetlands is affected by many man-made and natural factors, such as water supply, rainfall, and runoff [
34,
35,
36]. The environment in the study site of the lower Hailar River is complex and diverse with a variety of ecosystems, such as swampy wetlands, river meadows, scrub, and forest. This area is a transition area between the river system and the surrounding environment and is an important ecological transition zone in this watershed, which is the most active part of the ecosystem in terms of energy and material transfer, and transformation [
37,
38] and is very sensitive to external changes. These ecosystems are more dependent on runoff recharge from the Hailar River and periodic flood recharge, and their vegetation trends have both positive and negative possibilities. In areas with sufficient and stable water recharge, the vegetation gradually develops for the better, herbaceous plant communities gradually evolve into scrub woodland plant communities, bare soil areas remain stable, and biodiversity continues to grow. For areas lacking in water recharge, the speed of land desertification may exceed the speed of vegetation recovery, due to the influence of the adjacent.
Hulunbuir Sandy Land, thus forming a vicious cycle and leading to reverse ecosystem changes, such as vegetation changes to open sand habitat with sparse vegetation.
Human activities also affect vegetation changes. In this study, statistics of the Hailar Region for 1987, 1995, and 2019 were collected from Statistical Yearbooks.
Table 8 shows that when regional GDP increases at a relatively fast rate (more than 10 times that of the previous period), the population growth rate slows down, the growth rate of industrial production slows down, the growth rate of food production declines, and the growth rate of arable land increases by 266.99%. The increase in population and economic output leads to an increase in the scale of industrial and domestic water consumption in the river basin, resulting in an increase in the intensity of water resources development. These human development activities interact with the surrounding ecosystems and have become the main driving force in land expansion, agricultural intensification, and water resource exploitation in the Hailar area, and have profound impacts on the vegetation of floodplain wetlands.
In recent years, Chinese national and local governments have issued a series of wetland vegetation protection projects. For example, in 2003, the Chinese government approved the 2002–2030 National Wetland Protection Plan, which aims to restore natural wetlands and establish nature reserves [
39]. In 2011, Russia and China adopted the Russia-China Amur River Basin Transboundary Protection Zone Development Strategy to 2020, which announced that the protection of wetlands would be a top priority [
40]. Through a series of human interventions, wetlands have been restored and their vegetation protected. Therefore, in order to protect and restore wetland ecosystems, it is still necessary to sustainably manage natural wetlands [
39].
This research suggests that in the process of rapid urbanization, industrialization, and agricultural modernization, human beings should take greater responsibility for protecting well-developed wetlands and prevent their degradation. This requires, for example, better management of farmland irrigation, wetland development, and water resources.