Effectiveness of using global and national LDN indicators
This cycle of works represents a significant cluster (about 40%) of the total number of LDN-related publications, and reflect matters such as:
i. Validating indicators obtained from global datasets against national data;
iii. Variability factors for LDN global indicators;
iii. Possible alternative indicators for interpreting the dynamics of LDN global indicators;
iv. The necessity and feasibility of additional LDN indicators (or not using some global data), depending on local perspectives;
v. Ways of selecting a set of indicators considering regional and local peculiarities.
In relation to point (i) it was noted that three global LDN indicators do not much fit national data. At an individual farm level, the discrepancy between the results obtained using global data, data from detailed satellite imagery and ground-based observations can be as high as 20-30% (Belyaeva et al. 2019) for all three LDN indicators, in particular for SOC. (Lobkovskiy et al. 2022) indicate that the reason is that global LDN indicators reflect dynamics of certain parameters over a certain period of time, while the traditional national system is designed to assess the quality of land suitable for certain purposes (Table 1). In other words, the national system for land monitoring and assessing and existing relevant frameworks are based on sectoral methods of the assessment of land static qualities. This does not require data on comparable LD parameters obtained with a unified methodological platform. Moreover, land quality parameters that have the same names (for example, soil pollution) are often assessed differently in sectoral systems.
Studies related to point (ii) suggest that beyond physical impacts and other human activities, the variability of LDN global indicators is influenced by different combinations of factors such as climate change, natural fluctuations and successions in vegetation, geological and geomorphological processes associated with the development of river basins, and multidirectional trends in the natural development of landscapes in the postglacial epoch (Andreeva and Kust 2019). It has been demonstrated, that natural trends and processes are often more vigorous than anthropogenic ones, and without taking their background into account it is difficult to assess positive or negative land dynamics.
In relation to point (iii), it was demonstrated, that the assessment of land cover dynamics using the traditional national land accounting system does not fully correspond to the land cover classes adopted for estimating LDN (Table 1). In the Russian land cadaster land categories are distinguished according their intended use (forest fund, agricultural, industrial, etc.) and lands based on actual economic use. For example, the forest fund convers not only forested areas, but also includes arable lands, wetlands, roads, and other types of ‘land’. Statistical accounting for this system is complex and contradictory. In many cases the state of land cover in terms of ‘land categories’ remains invariant (especially for agricultural lands and forest fund), and the main changes (if any) concern ‘lands’, which can hardly be traced in state land statistics. Moreover, national statistics records the total actual areas, but do not record the transitions of ‘land categories’ or ‘lands’ from one to another. Therefore, without using space monitoring data, it is practically impossible to trace land transitions, and these methods are rarely used at the level of districts and individual farms.
The land productivity assessment through the use of vegetation indices obtained from the interpretation of multispectral satellite images is becoming a common place. However, for these assessment methods, there are also a number of difficulties associated with both the sectoral and biophysical features of different regions. For example, for drylands the NDVI derived from seasonal composites is more important than the annual average. (Zolotokrylin, Titkova, and Cherenkova 2020) discovered that moisture changes in early summer affect the state of vegetation more significantly than those of the full vegetative season. In (RSF 2020) report it was demonstrated that for farmland the NDVI assessments is of little use for irregular crop rotations, since the spectral reflectance of different crops does not repeat in a multi-year cycle. Even for regular crop rotations individual recognition algorithms are required for cycles of different terms. Data on the key crops yield can serve as an alternative, however, they have significant fluctuations depending on the weather conditions of specific years and on applied agricultural techniques and crop varieties, making it difficult to trace the patterns of their dynamics. For boreal forests, it was determined that the increase in biomass corresponded to the NDVI dynamics, but the growth rate decreases in long term, when young forests pass into mature ones. The productivity of a mature forest, calculated by the NDVI, is usually less than the same for young forest, although this does not reflect degradation of forest land. Alternative indicators could be: timber stock, proportion of forest cover, total growth. However, the biggest challenge is NDVI interpretation for abandoned farmland overgrown with weeds, shrubs and trees. The ”greenness” of the surface of such lands is in the majority of cases higher than on the adjacent farmland, but from an economic viewpoint these lands are more degraded than the actual arable land.
Estimation of SOC change is a practically insoluble problem at country level: global and national datasets do not correspond each other (National Report 2019). The SOC dynamics in most areas have not been regularly monitored by field observations since 80-s. For those regions where it remains possible (only for farmlands), the SOC dynamics rather than any long-term trends illustrates a short-term perspective of the economic possibilities of farms for the use of organic fertilizers and available agricultural techniques (Bezuglova, Nazarenko, and Ilyinskaya 2020).
For the purposes of local assessment of agricultural and forest enterprises, and in the absence of reliable data, it was demonstrated that individual software tools cannot yet adequately assess carbon exchange processes, especially in steppe and forest-steppe landscapes. However, correctly selected ensembles consisting of simulation models of C-balance and C-calculators supported by field data, are able to successfully solve such problems. (Karelin and Tsymbarovich 2022) determined that if this condition is met, then the net carbon balance of a certain area can serve as an alternative to SOC. The possibilities of replacing the SOC layer in the Trend.Earth model with a carbon balance layer (calculated using Ex-Act tool (FAO, 2021)) for mapping LDN at the individual farm level have been identified (RSF 2020). However, for boreal regions, the situation is more complicated: (Ptichnikov et al. 2019) showed that the net carbon balance cannot be applied as an independent LDN indicator, since it does not consider changes in biodiversity and primary productivity. The simulation models for natural forest dynamics through minimization of forest felling at sites with fireless types of succession accumulating maximum of dead phytomass, could be more applicable.
Point (iv) is related to using additional and alternative indicators, which would allow to more accurately indicate degradation trends in achieving LDN. In this aspect, the soil erosion indicator is the most important for Russia. It can be interpreted in terms of different measures. It is asserted that in Russia the total area of eroded lands and those under the risk of erosion make up more than 50% of agricultural lands. However, recent studies indicate a decrease in erosion rate and in the total area of eroded land during the last 30-40 years as a result of abandonment of farmlands and subsequent overgrowth by natural vegetation. Climate change has resulted in a decrease of the depth of soil freezing and the flow of spring runoff, which in turn leads to a decrease in soil erosion. (Tsymbarovich et al. 2020) demonstrate the possibility of using the soil erosion indicator as an important complement to the three global LDN indicators both at the country and local level. At a national and subnational level, soil erosion neutrality can be interpreted through two indices “Rate of soil loss” (ton ha-1 yr-1) and “Total soil loss” (ton yr-1) when using a “one-out-all-out” approach.
(Karelin and Tsymbarovich 2022) demonstrated that soil water erosion indicators, (in particular length&steepness (LS) factor; the erodibility potential) can also be used for indirect interpretation of SOC, nitrogen and water dynamics, as well as in reduced presence and altered activity of soil microbiota. The authors showed that the LS-factor calculated on the basis of remote sensing data is applicable for the evaluation of erosion hazards, as well as for prediction of carbon content and other related physical, chemical, and biological indicators of arable Haplic Chernozems on a large spatial scale. The spectral characteristics of the soil surface obtained from remote sensing data are less applicable for these purposes.
The indicator of aridity is the next most important indicator for LDN assessment (Kuderina et al. 2020). At a local level, this can be interpreted through measures of atmospheric precipitation (including snow accumulation) and soil moisture in different seasons, which integrally assess the ability to retain moisture during the growing season and thereby determine the heterogeneity of the soil cover and landscape dynamics. A series of drought indicators that ultimately reflect soil degradation through the availability of moisture and nutrients to plants include characteristics such as soil water holding capacity (Shcherba et al. 2016) and compaction (Sorokin and Kust 2018), (Tsymbarovich et al. 2020). Similar approaches are demonstrated in the works of Belarusian researchers (Yatsukhno and Davydik 2018). Such drought indicators are primarily relevant for local LDN estimates.
At a subnational level, an increase in aridity in certain regions, the emergence and growth of “islands of climatic desertification” can also serve as an indicator of land destabilization and degradation. Such impacts of overgrazing were demonstrated in Mongolia and in southeast of European Russia (Zolotokrylin 2019). The area of local lakes can serve as an indirect integral indicator of aridity, climate change and the water availability. For example, the dynamics of feeding mechanisms of lake systems in south-western Siberia and Transcaucasia indicate either a growing aridity (Shaporenko and Abdurashidov 2021) or very moderate and decreased anthropogenic impact (Chernykh et al. 2022). The frequency of forest and steppe fires (Shinkarenko et al. 2022) can also be considered as an additional LDN indicator at national and subnational levels. Importance of drought indicators for monitoring LD processes is reflected in (National report 2021).
Indicators of soil salinity proposed by (Chernousenko 2021), agro-depletion (decrease in the content of basic nutrients), alkalinity (Tsvetnov et al. 2020, Tsvetnov et al. 2021a, Makarov et al. 2021a) can be considered for LDN at different levels. This requires a well-established monitoring system, which is currently under development (Unified system 2021). For forest areas additional LDN indicators were proposed: level of biological diversity (indicated by species diversity of trees and shrubs) (Ptichnikov and Martynyuk 2020), and the types of recovery dynamics (also accompanied by changes in species diversity) (Ptichnikov et al. 2019).
(Zolotov et al. 2020), (Kust et al. 2021) have emphasized the possibility of using the state of protected areas (PA, area in the region and internal stability) to characterize changes in land cover. For example, the total area of PAs in the Altai region due to the establishment of new PAs and expansion of existing PAs will increase from 795.55 thousand hectares or 4.74% of the region’s territory to 1616.6 thousand hectares (9.6% of the region’s territory) and preserve the most significant natural complexes. A twofold increase in the area of protected areas will undoubtedly contribute to the achievement of LDN.
The last point (v) is focused on selecting a set of indicators taking into consideration regional and local characteristics (Andreeva and Telnova 2018, Kust et al. 2019, 2018a, Lobkovskiy 2020, Lobkovskiy et al. 2020a, Lobkovskiy et al. 2018). It was observed that in the current Russian land registration system and statistics there were no indicators corresponding to LDN indicators used for monitoring LD, except for in the case of agricultural lands. In order to integrate land quality indicators used in Russia into the global system, it has been proposed to streamline various sectoral data through a common consistent LDN-based “superstructure” for different sectoral systems. A conceptual scheme of a hierarchical structure for a system of LD indicators in Russia has been developed (Lobkovskiy et al. 2022), including, on one side, dynamic indicators characterizing the achievement of LDN, and on the other side, state indicators characterizing the quality of land (in terms of risks and degradation results) in relation to their respective categories (Figure 6).
Using these approaches, a preliminary analysis and selection of indicators for those operating in different sectors was carried out, in order to integrate into the LDN-based global assessment system. The possible indicators are: area of eroded lands, forest cover, average yield, etc. Further ways of integrating national and international systems for assessing LD are as follows: (i) development of a unified national list of indicators and measures for lands of different categories and types of land use; (ii) use of common qualitative scales of indicators while varying the quantitative values of their measurements (by region, industry, land category, etc.); c) use of land quality scales comparative to either best (maximization of indicators) and/or worst (minimization of indicators) sites in a certain locality/region.