I. INTRODUCTION

Land degradation (also labelled 'desertification' when in drylands), restoration of degraded lands, and invasive alien species, have each been spotlighted in thematic assessments of the IPBES: the 2018 "Land Degradation and Restoration" and the 2021"Invasive Alien Species"). Land degradation is often expressed as long-term reduction or loss of biological productivity which is of value to humans; invasive alien species are species introduced into locations off their natural distribution range, where they then impact the new area's biodiversity and ecosystem services. The terms "land degradation" and "invasive alien species" both are colored by negative overtones ('degradation', 'invasion'), and they are habitually presented as "serious and rapidly growing threats" to livelihood security (IPBES 2018). Although assessed and often attended separately, these two entities often interact, such that the above threat is often exacerbated. This interaction, however, is not sufficiently understood and deserves further research (e.g., Richardson and Van Wilgen, 2004); Shackleton et al., 2014).

This review paper presents a case study largely based on secondary data analysis of this land degradation-invasive alien interaction, in an Indian arid grassland ecosystem invaded by an alien plant, and highlights its half-century vicissitudes in a social-ecological system, i.e., the Banni Grassland degradation and its restoration, both driven by the same invasive alien Prosopis juliflora (Pj in this study).

II. BANNI GRASSLAND SOCIO-ECOSYSTEM

a) The Banni grassland

The Banni (meaning 'newly created') grassland is a cloud-shaped area extending 2,618 km ${}^2$ (GUIDE, 2007) in the north-western area of India's state of Gujarat (230 19' - 230 52' N). It is bounded by a huge marshy salt desert (Great Rann of Kachchh) at its northern border and a smaller one to its south (Fig. 1). It is regarded as the only remaining single natural vast stretch of grassland in the Indian sub-continent, and, despite the inherent salinity of its alluvial sandy lands, it is seen as Asia's finest Grassland (GUIDE, 1998).

Banni soil is saline due to saline groundwater surfacing through capillary movement, and then evaporating under the arid dryland climate (335 mm average annual rainfall, and frequent droughts (26 from 1932-2008), thus leaving salt deposits on the topsoil (Kumar et al, 2011). The topsoil also has low water permeability (0 to 0.13 cm/hr.) due to its high proportion of silt and clay, causing flooding and water logging (Singh and Kar, 2001). The level of soil salinity varies spatio-temporally, depending on the specific site's distance from the bordering marshy saline desert areas, and on the frequency and intensity of the seasonal leaching effect of freshwater flow from seven rivers that include Banni in their catchment areas. Together, the areas of highly saline (over 15.0 M mhos/cm) topsoil are spread over 51% of Banni (Soil Survey Division 1986; GUIDE, 1998). In spite of these harsh abiotic environmental features, Banni is endowed with relatively rich biodiversity (91 plant species, 81% of them common, GUIDE, 2011) typical to productive grassland ecosystems, and since 1955 it has been credited with the legal status of an Indian "protected forest".

b) The Banni

Given records as of the ${18}^{\text{th}}$ century Banni people's roots are in Banni's adjacent Sindh desert, where they practiced a pastoral lifestyle based on livestock grazing, and temporarily settle in Banni. In the early ${19}^{\text{th}}$ century, these family/clan settlements gradually increased in permanency until they qualified as "villages", even though they maintained the traditional Banni–Sindh nomadism. However, in the late 19 ${}^{th}$ and early 20 ${}^{th}$ century, the ruler of the principality in which Banni is embedded (currently Kachchh district, see Fig. 1) declared the area a "Reserved Grassland", in which settlement is forbidden, thereby reinforcing the Banni–Sindh desert nomadism.

Following India's independence in 1947, the Indian state of Gujarat allowed grazing in Banni, but imposed livestock-head taxing, abolished in the mid-1960s. This, together with the India-Pakistan partition, deterred the Banni–Sindh nomadism. These policy-social-cultural processes increased livestock and human populations and encouraged permanence in an increasing number of small residential settlements (48 "villages" by 2015). Each village comprised a cluster of small family/clan houses positioned next to a local water source (Trivedi, 1965) and surrounded by grassland areas where free-ranging livestock forage, such that the entire Banni grassland became available for grazing.

The India-Pakistan partition of 1947 and the abolishment of the livestock tax in the mid-1960s led to further reduction of the Banni–Sindh nomadism. Gradually the pastoral livestock owners settled in the permanent residential settlements, with livestock (primarily cattle and seconded by buffalo) the mainstay of their livelihood, while sheep and goats were common in number but not that economically significant (for this paper 'livestock' is used for cattle and buffalo only). The overall grazing condition of the Banni was fair; the livestock depended solely on the forage provisioning service of the Banni grassland ecosystem (green forage production of 465 tons/grassland ha), with no overgrazing and no added or stored fodder, except in drought years, when the herds migrated to less arid areas in southern Gujarat and its southern border state of Saurashtra.

The Banni grassland herders were unique in that their grass-dependent livestock products were used not only for their own consumption and subsistence, but also engaged in income generation, primarily by sale of bullocks for traction, land ploughing and levelling. Thus, the Banni people, locally called Maldharis (cattle breeders), are known not only as livestock herders, but also for their skills in breeding superior cattle and buffalo breeds, including the Kankrej breed (Fig. 2) – the heaviest cattle breed of India, highly adapted to extreme and prolonged drought conditions. Unlike a livelihood based on livestock products, whose success is subject to Banni's climate variability, bull breeding is a drought-proof adaptation; in rainy years the bulls are reared, and in drought years they are sold. The Maldharis's traction bulls are regarded as one of the best breeds in India and were sold in most parts of Gujarat, as well as in neighbouring states. In 2011 for example, a price of a Banni bull was about five times higher than that of a Kankrej cow, and its yearly average sale of ca 15,000 bulls jointly fetched $15,385to$18,462, what supported 6,000-10,000 villagers (Mahajan&Bharwada, 2011).

Furthermore, the Kankrej breed has a high milk yield. Traditionally, the milk's significance is not in its market value (and hence should not be sold), but rather for its importance in nourishing the calves, thereby maintaining the quality of the breed (Trivedi, 1965). On the other hand, milk production of the Khunni buffalo – a local indigenous breed (Fig. 3) that is also highly adapted to the Banni environment – is among the heaviest and highest milk productive breed in India.

Thus, although the buffalo is used mostly for the herders' subsistence, it also generates income through the production of ghee, a buffalo milk product of high fat concentration, used not only by the herders, but also sold locally.

III. THE ALIEN INVASIVE PROSOPIS JULIFLORA

a) The

Prosopis juliflora is an evergreen thorny leguminous shrub/small tree of 3-15 m tall but usually up to about 5 m tall, and the 20 to 30 cm long pods are in clusters of 2 to 5 at the ends of the spreading branches. It is native to the arid and semi-arid zones of the subtropical and tropical regions of northern South America, southern Central America and the Caribbean (Fig. 4).

Related species, Prosopis pallida is also native to these regions, whose center of distribution is in Peru, whereas that of P. Juliflora is in Ecuador and Columbia, and a hybrid zone of the two occurs in northern Peru, Ecuador and southern Colombia. Due to their similarity and hybridization, they have often been misidentified to the extent that an in-depth review (Pasiecznik et al., 2001) recommends attending a "P. juliflora - P. pallida complex" rather than each separately. Given the specific source of the plants introduced to Banni grassland, some characteristic of the invading plants, and the prevailing naming by stakeholders, for the purpose of this paper the Banni grassland invader alien is labelled as Prosopis juliflora (hereafter cited as Pj).

In its native range $P_j$ occurs in thickets (Central America and the Caribbean, Pasiecznik et al., 2001) or mixed with other species in scrub vegetation (southernmost Ecuador [Neill, 2015]) and Galapagos (Wiggins and Porter, 1971). As of the 1800s $P_j$ populations in their native ranges have been often the source of $P_j$ "weedy invasions" of neighbouring areas. These population spreads were triggered by deforestation of invaded areas converted to rangelands and cultivated lands (Pasiecznik et al., 2001). It is likely that these rather local weedy invasions triggered by land use changes demonstrate the qualities of $P_j$ in withstanding a wide range of environmental conditions.

and in rapid production of fuel wood and timber, attributes that encouraged the intentional introductions of $P_j$ to many countries as of the late19 ${}^{th}$ Century (Senegal, year 1822), and further introductions to the Americas (2 countries), Oceania (3 countries), Africa (25 countries) and Asia (18 countries), mostly motivated by fuel wood shortage (Mwangi and Swallow, 2005). In most countries to which $P_j$ had been introduced, thus millions of hectares, mainly of rangeland have already been, and are still being invaded (Pasiecznik, 1999) such that in 2004 $P_j$ ranked as "one of the world's top100 worse invasive species", by the Invasive Species Specialist Group of the IUCN. Major areas already invaded are in Argentina, the Sahel countries, Kenya, Australia and in most of the arid and semiarid regions of India.

b) Pj

Pj first arrived to India from Mexico to Sindh, an arid region north of Banni (currently in south Pakistan) in 1857 (Reddy, 1978; Muthana and Arora, 1983). Later introductions to India have been: from Jamaica to the State of Andhra Pradesh (far south of Banni) in 1877 (Raizda and Chatterji, 1954; Reddy, 1978); from Peru, Mexico, Argentina and Uruguay to many parts of India in 1915-1916 (Burkart, 1976); from Mexico, introduced for stabilizing sand dunes in Rajasthan's Thar Desert during the 1930s, where it was officially declared a "royal plant" to be protected (Muthana and Arora, 1983); and from Mexico and Africa, for planting close to the city of Morbi in the Gujarat region of Kachchh, ~100 km from Banni in the 1930s, for "fighting desertification" (TNN, 2004). Thus, by 2000, Pj covered nearly 40% of India (Tewari et al., 2000) and, with 43 million trees, became the dominant tree of Gujarat (TNN, 2004).

But not until 1960s was $Pj$ introduced to Banni by the Gujarat State Forest Department, an introduction driven by a trend of intensifying soil salinization beyond the values inherent to the Banni area. The increased salinity was triggered by two policy interventions: 1) check-damming of rivers (see 2.1.), which reduced the amount of salt leached from the soil (and also reduced the river-deposited supply of nutrients); and 2) new road construction for national security purposes causing diversion of seawater flow producing seawater-tidal inundation of the "Greater Rann of Kachchh," which increased the windborne salt from the Rann, thus affecting 12 villages in Banni along its northern border with the Rann (GUIDE, 1998). The increased soil salinity resulting from these two policy-driven externalities led to lowered productivity of the grassland, both quantitatively and qualitatively - an example of land degradation, i.e., "an anthropogenically driven reduction of biological productivity" (IPBES, 2018) on which most services of the Banni ecosystem and their benefits to Banni people, fully depend.

This trend of reduced productivity driven by increasing soil salinity has been visibly severe since the early 1960s, as noted in the third Five Year (1960-1965) Plan of the Gujarat State Forest Department, which described "disappearing vegetation" and "conditions conducive to further desert formation" (Government of India, 1960). The Department, impressed by Pj's higher salinity and drought tolerance (relative to that of several other species it had tested), set out to plant from 1960/1961-1965, a Pj shelter belt in a 315 km ${}^2$ area on the northern edges of Banni (Fig. 1). The area, comprising 12% of the Banni, was intended to function as a wind breaker that would reduce the salt particles blowing in from the Greater Rann of Kachchh into Banni (GUIDE, 1998; Tewari et al., 2000; Bharwada and Mahajan, 2006; GUIDE, 2011), thus protecting Banni's forage provision and supporting the herders' livelihood.

c) Pj invasion adaptations

Since its arrival in Banni in the 1960s, Pj has swiftly breached all post-introduction invasion barriers (survival, reproduction and spread), to become a "self-sustaining population with individuals surviving and reproducing also at significant distance from the original point of introduction", thus earning the designation "invasive alien" (Blackburn et al., 2011). This can be attributed in part to the plant's dryland-specific invasiveness traits. Itself a native of drylands, $Pj$ is adapted to live with relatively limited resources of soil moisture and nutrients and with the constraints of high soil salinity and alkalinity. Other major invasion adaptation traits are $Pj$ 's life-history variables of reproduction and survival. Flowering twice a year (starting from the first year of life, Pasiecznik et al., 2001) and producing 31,700 relatively large (13x0.75x0.54 cm) pods/tree/year at maturity (Felker, 1979)that cumulatively contain 805,000 relatively light-weight (0.03 gr) seeds/tree/year (Fenner, 1985; Shukla et al. 1984; Silva, 1986) ${}^1$, account for $Pj$ 's high reproductive rate. This, combined with the plant's high survival rate at all stages -- seedling, sapling and adult tree (all of which are inedible to wild and domestic herbivores), contribute to its high population growth rate in Banni.

Moreover, another important trait, which accounts for the plant's extensive spatial spread, is its high germination rate (much higher than of Banni indigenous Prosopis cineraria, El-Keblawy and Al-Rawai 2006) and those in deeper soil layers remain dormant, acting as a reserve during times of scarce rainfall. In addition, high rate of $Pj$ seeds become seedlings and survive to adulthood with a lengthy survival probability. These survival successes can largely be attributed to a favourable set of circumstances: To begin, unlike its inedible foliage, $Pj$ pods' pulp is highly nutritional in carbohydrates, proteins, fats and minerals, and is thus palatable to and edible by livestock. Second, the season of $Pj$ pods' maturation, shedding and abundance (41 tons of pods/ha of 2,053 trees/year, Gavali & Lakhmapurkar, 2009) occurs when most grass and herb species of the grassland have already been grazed by Banni livestock or are too dry for consumption, and so the livestock (mostly buffalo, which are more numerous and better adapted to pod consumption than cattle) consume these seasonally shed pods, along with their packed seeds. Finally, the thick seed coats are disintegrated by the animals' digestive tract acids – an essential pre-treatment for germination – and are excreted after having been well fertilized with adequate moisture from the animals' dung. Since Banni's livestock are free-ranging, and since a livestock animal defecate about 14 times/day (Weeda, 1967; GUIDE and GSFD, 2010 ${}^2$ ), most $Pj$ seeds are randomly dispersed in a high diversity of germination sites at considerable distances from the parent plant. This accounts for the plant's high seedling survival rate and consequent extensive spread in Banni.

$Pj$ 's invasion adaptations traits for population growth and spread alone, however, are not sufficient for the plant to compete successfully with the diverse indigenous plants of Banni's thriving ecosystem. $Pj$ 'snative lands (see III a) are shrubland/woodland ecosystems in which $Pj$ rarely dominates and is sometimes totally outcompeted, unless other tree species in the area are removed by man. But Banni is a grassland ecosystem whose salinity excludes potential tree competitors. Thus, introduced into Banni, $Pj$ did not encounter dominant native tree species, but rather the grassland's 30 common native grass and 4 common native sedge species (GUIDE, 2011) with which it would compete for the scarce Banni resources with two competitive advantage traits: 1) addressing water scarcity, the plant's foliage absorbs moisture from dew and from light rains, but more important is its dual root system, with both surface lateral roots and depth vertical roots, a root system instrumental in outcompeting both topsoil-rooted annual plants and deep-rooted perennial grass species (Mooney et al., 1977; Pasiecznik et al., 2001); and 2) addressing soil nutrient deficiency, $Pj$ interacts with several strains of salinity- and drought-tolerant Rhizobium bacteria and several mycorrhizal fungi in atmospheric nitrogen fixation (Pasiecznik et al., 2001).

These resource competition advantages alone may not have enabled $Pj$ 's high population growth and spread in the saline Banni grasslands. $Pj$ was introduced to Banni with the goal of slowing the grassland's increasing salinisation process (see 3.2) by curtailing the airborne salt input with a wind breaker tree, which itself has a high salinity tolerance (up to 36,000 mg/l NaCl, [Felker et al., 1981] and 6,000 mg/l salinity with no reduction in growth [Harsh and Tewari, 2009]). Paradoxically, this trait gave $Pj$ an unexpected competitive edge over the indigenous plant species inhabiting the grassland's spatially diverse salinity levels, most of which have a lower salinity tolerance than $Pj$.

An example of the spatial variability of salinity levels of Banni lands is illustrated by a survey of nine sites covering an east-west transect of Banni, yielding a salinity range of 0.43-7.37 ppt, with a mean of $4.25\pm 2.29$ ppt, thus having a coefficient of variation of 0.5 (Fig. 5, salinity range, Gavali & Lakhmapurkar

2009). Consequently, the indigenous plant communities that inhabit Banni lands of diverse salinity levels also differ in their salinity tolerance levels. When introduced to Banni, $P_j$ 's density (trees/ha) was negatively linked to the salinity level of the area it invaded (Gavali & Lakhmapurkar 2009), as well as to the number of its potential indigenous competitors: high quality palatable grass species and indigenous tree cover tolerable to "very low and "low" salinity categories; salinity sensitive nutritious grass species give way to more resilient but less nutritious species tolerate soils of "high" and "moderate" categories, and the "extreme" salinity excludes all palatable grass species (GUIDE, 1998). Thus, the higher the salinity of the invaded land, the smaller the number of species with which $P_j$ invaders would need to engage and outcompete (Table 1). Thus, under Banni's environmental conditions, high salinity tolerance qualifies as $P_j$ 's critical invasion adaptation trait.

Salinity dS/m & Mmhos/cm	Category	#Pj/ha	Trees - Acacia nitotica, Prosopis cineraria Shrubs - Capparis, Salvadora	Dense, high quality palatable grass Sporobolus, Dichnithium, Cenchrus, Sesmostachya	Low quality palatable, Nutritious grass/salt grass, Aeluropus logopoides	Sparse, mostly non palatable, Low nutritious grass Eurochondra setulosus
<1	Very low	2580
<3	Low	2440
<8	Moderate	2220
<15	High	970
>15	Extreme/Very high	Sparse

Added to its advantage in resource competition, apparently intensified by its high salinity tolerance, $Pj$ also employs interference competition, as some of its leaf, fruit and seed chemicals function as allelochemicals that negatively impact the growth and seedling germination of neighbouring plants (Noor et al. 1995). Furthermore, $Pj$ reduces its vulnerability to insect herbivory and to large herbivores' browsing by investing in yet other leaf chemicals, mainly alkaloids and tannins that deter insects and make leaves non-palatable to mammal herbivores, whose browsing is also impaired by $Pj$ 's 0.5-4 cm long thorns (Pasiecznik et al., 2001). Jointly, these attributes endow $Pj$ with a high potential for competitively monopolizing land resources, leading to rapid attainment of stable population size at the carrying capacity of the invaded Banni grassland ecosystem. With this, $Pj$ 's introduction to Banni grassland's northern border (Fig. 1) in the early 1960s triggered a sequence of ecological, social, economic, and policy changes in Banni and beyond.

IV. THE INVASION AND ITS IMPACT

a) Pj invasion following introduction

Upon its introduction into Banni, Pj encountered two land-cover challenges – competing with the dense grass species communities and coping with the vegetation-barren, high salinity areas. Throughout the first post-introduction decade, Pj sparsely invaded the former and largely avoided the latter (Fig 6a).

• (a) Image of 1979, with northern section masked; *(b) Image of 2009, black diagonal line marks the 1979 southern border of the areas masked. The Banni imageries of the years following that of 1979 are not masked.

There are no records from the initial invasion stage during the first decade following the 1960s planting. Unfortunately, the $Pj$ planting area (Fig. 6a) was masked for security reasons (see 2a) in the first satellite image of Banni, obtained in 1979. Nevertheless, $6.6\mathrm{\%}$ of the non-masked area, added to the recorded size of the planting area, suggested that by 1979 – a decade after its initial introduction – $Pj$ was already established over nearly $20\mathrm{\%}$ of the Banni grassland. Thereafter, dense cover of $Pj$ expanded at an average rate of 2,670 ha/year through the second decade (Jadhav et al., 1992), then gradually lessened to 1,664 ha/year by the fourth decade until, in 2009, it densely covered $30.7\mathrm{\%}$ of Banni (Fig. 6b, GUIDE, 2011), and sparsely covered another $24\mathrm{\%}$.

This pattern of invasion illustrates how an area is at first sparsely invaded by young plants, and as they continue to grow and increase in density in that area, other areas become sparsely covered by new invaders. Thus, in the early stages of the invasion in the early 1980s, the areas of sparse distribution were more extensive than those of dominant distribution. But by 2009, Pj had attained dominance in most, but not all invaded areas, and the extent of Banni area covered with sparse Pj remained stable (Fig. 6c). Forty years after the end of its introduction period (from 1960 - 1970), Pj had invaded 55% of Banni (Fig 6b), transforming significant areas of grassland to woodland, and thus reliably deserving the title "transformer" invasive, i.e., an "invasive plant that changes the character, condition, form, or nature of ecosystems over substantial areas" (Richardson et al., 2000).

However, in many areas of Banni, $P_j$ distribution remained sparse (Fig. 6b, 6c). This is because the spatial invasion pattern was not that of a front advancing from the site of introduction, but rather of a patchy random distribution (Fig. 6b). $Pj$ densely covered a substantial area in the heart of western Banni emerging from the introduction site, and two additional dense patches, at the western and southeastern edges of the grassland, $\sim 5$ km and $\sim 20$ km distant, respectively, from the larger area. Thus, the 2009 satellite image of Banni (Fig 6b) presents what appears to be a random pattern of $Pj$ and grass patches embedded in a vegetation-barren, saline land cover. In fact, however, this "random" mix of patches of varying density, stable sparse patches, and a reduced invasion rate with time, mostly reflects the spatial variability of Banni's soil salinity – the major driver of Banni ecosystem change that affects both the invasive $Pj$, the indigenous grass cover, and their mutual interaction (see 3.c).

(Figure 6c): 1980-2009 invasion trends

Concurrent with $Pj$ 's introduction in the early 1960s and its initial spread through the 1970s, Banni's inherent soil salinity was also increasing. From the first river damming in 1955 to the seventh in 1981, the freshwater flow into Banni continuously decreased, resulting in increased soil salinity year by year. This, with the inundations following the 1965 road construction (see 3b), jointly initiated an emergent salinity surge, a process of transforming low-salinity lands to higher salinity within the <1 to >15 Mmhos/cm salinity range prior to the simultaneous $Pj$ and salinity upsurges. Furthermore, the salinization trend, concurrent with the $Pj$ invasion trend, has not been stable, but rather has been affected by intense rainfall oscillations. For example, from 1975 to 2014, annual rainfall ranged from 0 – 896 mm, with an average of 375 mm (CV=0.65), and there were 13 very severe droughts ${}^3$; the high rainfalls generated floods that reduced topsoil salinity, and the high frequency droughts increased it.

For example, from 1980-1992, areas of "extreme" (highest category) salinity, $Pj$ and grass land covers all showed increasing trends. While the size of the highly saline lands fluctuated intensely with a 24% difference between lowest and highest area values, there was only a 12% and 8% difference in $Pj$ and grass land covers, respectively. These oscillations were followed by intense rainfall oscillations (0-657 mm/year) within this same period, which apparently affected the $Pj$ and grass land cover size oscillations. The plant and rainfall oscillations were not fully concurrent, apparently due to the plants' delayed response to climate variabilities, often expressed by either flood or drought events (Fig. 7a).

Hence, Banni's surge in soil salinity may have acted as double-edged sword – decreasing the indigenous grass land cover and empowering the competitive edge of the alien $Pj$, thus advancing its spread. Fig. 7a illustrates the relative contribution of the salinity surge and the $Pj$ invasion surge to the grass distress in Banni. The barren area of extreme salinity continued to expand throughout the last decade of the ${20}^{\text{th}}$ century and the first decade of the ${21}^{\text{st}}$ century, from $715{\text{Km}}^2$ in 1989 to $1,134{\text{Km}}^2$ in 2009. This expansion coincided with a loss in productive grassland area of $798{\text{km}}^2$, while the extreme salinity soil, i.e., land that lost its grass cover, "gained" only $419{\text{km}}^2$ (Table 2). This may indicate that the other $379{\text{km}}^2$ of lost grass-covered area was transformed to Pj shrubland/woodland land cover. And indeed, during the same period, Pj range increased by 392km ${}^2$.

These data suggest that the loss of the grassland cover in Banni can be attributed approximately evenly 379 and 392 Km ${}^2$, to the increasing salinity and to its competitive exclusion by Pj (Table 2). The mechanism for these land cover exchange may be straightforward; i.e., the increased salinity paved the way for Pj to easily invade areas in which the grass cover had been reduced by the salinity surge. However, the initial Pj surge, from 26.7 Km ${}^2$ /year of the 1980-1988 period, to 54 Km ${}^2$ /year of 1988-1999 (or 48 Km ${}^2$ /year of period 1988-1998, Jadhav et al., 1998; Jadhav et al., 1992) has not been maintained, but reduced to 16.7 Km ${}^2$ in the 1999-2009 following period.

	1989 area in Km2	2009 area in Km2	Difference area in Km2	Km2 of grassland transformed to Pj cover*
Pj	412	804	+392 ≈ 379
Grass	1429	631	-798
Saline	715	1134	+419

Nevertheless, most grass species, among them many palatable to livestock, are not totally outcompeted by $Pj$, as evidenced by their cover of spaces between neighbouring $Pj$ trunks. Even though $Pj$ litter scattered around the trunk is toxic, most of the grasses covering these spaces are out of the litter's toxicity range. Yet, as $Pj$ plants become trees, they constitute a physical obstacle to livestock mobility such that the grass plants that evaded $Pj$ competition and toxicity remain inaccessible to the livestock. Thus, the biological productivity, and hence the biomass, of $Pj$ cover are higher than that of the grass land cover that it replaced, and even though this replacement allowed grass plants to survive, the economic biological productivity of the lands invaded by $Pj$ declined significantly.

*The 798 kilometers of lost grass cover, minus the 419 increase in saline area kilometers, equals 379 kilometers, approximately equal to the 392kilometers gain in $Pj$ land coverarea.

b) Pj driving ecosystem change, a contextual land degradation

The process of invasion of alien species, $Pj$ included, frequently drives a process of ecosystem change that is often detrimental to the local people whose livelihoods depend on the services the ecosystem provides through its specific biodiversity components. Prior to $Pj$ 's introduction, the Banni grassland's major ecosystem service provisioned forage for livestock – the mainstay of the Banni people – with a diversity of livestock-edible nutritious and palatable grass species. Banni's grassland ecosystem also contained woody plants typical to arid saline drylands: 24 shrub and 17 tree species, including Suaeda, Ziziphus, Tamarix, Acacia and even a Prosopis species – P. cineraria. But Banni was not a woodland ecosystem invaded by a woodland plant. Rather, $Pj$ invaded a primarily grassland ecosystem, whose land was covered mostly by coarse and low perennial plant species - 6 sedge, 89 herb, and 37 grass species (GUIDE, 2011).

Most of these species are palatable (Patel et al., 2012) but they accounted for only one-third of Banni's land-cover biomass (ton/ha); 10 grass species comprised 76%, and 8 herb species made up 33% (GUIDE, 2004). Furthermore, in 1965, the three dominant palatable grass species (Dichanthium annulatum, Cynodon dactylon and Eleusine flagellifera) constituted 32% of Banni's plant cover, and one non-palatable ${}^4$ grass species (Desmostachya bipinnata) accounted for 68% (Trivedi, 1965). Thus, prior to Pj's invasion, even though much of Banni's grass land cover was not livestock-palatable, it nevertheless nicely supported the livelihood of its residents.

However, a decline in the grass and sedge species richness following $Pj$ 'sinvasion was detected when a 2012 survey did not record 12 of 41 grassland species registered at the same site 30 years earlier, in 1982 (Pandya and Sidha, 1982, cited by Singh and Kumar, 1998). Furthermore, the grasland economic productivity, expressed in plant biomass annual yield (tons/ha of livestock-palatble species harvested from forage-protected plots), also declined, from 4.65 ton/ha in 1961 to 0.48 in 2004 (Trivedi, 1965; GUIDE, 2004; GUIDE 1998). The data clearly show a decline in grass yield (Fig. 8), even though four of the eight years surveyed had high rainfall (Saxena, 1993). This productivity declines apparently occurred in lands not yet densely invaded by $Pj$, but that were undergoing a concurrent salinization process, transforming low-salinity lands to high- or moderate salinity (Table 1, Fig. 6c).

This can be attributed to the salinization process that reduced the grass density in these lands, facilitated by $Pj$ invasion to these lands, but only in its initial sparse stage (Fig. 6c).

The reduced biological productivity of an economic value that detrimentally affects its users (in this case, the Banni people), and is driven by human interventions (in this case, river damming and $P_j$ introduction) falls within the definition of "land degradation": "land degradation means reduction or loss in...arid areas... of the biological or economic productivity of... range, pasture... resulting from processes... arising from human activities... such as...long term loss of natural vegetation"; and when occurring in arid areas like Banni, it falls within the "desertification" definition: "Desertification means land degradation in... arid areas resulting from... human activities..." (Article 1, Use of Terms), United Nations Convention to Combat Desertification, UNCCD, no date).

${}^1$ Bhirandiyara village (Trivedi, 1965) ${}^2$ GUIDE, 2004

Data collected by Banni Development Agency (GUIDE, 1998)

4 values in red – ton/ha in a rainy year

Figure 8: Annual (mostly palatable) grass species productivity in tonnage per hectare

The Banni case can be also addressed in the context of land degradation/desertification process that is not attributed to climate variation and/or to land users' mismanagement, but to invasion of alien species.

The concurrent processes of soil salinization and $P_j$ invasion that triggered the desertification and land degradation of Banni lands through gradual but steady transformation of its grassland ecosystem to one of shrub/woodland often leads to an ecosystem-service provision decline, or to a services trade-off. In Banni, not only did the livestock forage service decline, but its benefit – the income from bull/bullock sales – was challenged without a service trade-off. Demand for traction bulls dramatically declined and the grassland forage provision gradually lost its users/customers. As a result, the sale of Banni bulls for traction drastically declined.

c) Exogenously-driven vicissitudes of Banni ecosystem and community

The sale of Banni bulls for traction drastically declined when, by the beginning of the 1990s, about one-fourth of Banni was already under dense Pj cover (Fig 7a) and increasing salinity encroaching upon the grassland caused the loss of more than a quarter of Banni's pre-Pj invasion grassland yield (Fig. 5c, Fig. 8). These two processes coincided with a decrease in number of draught animals (80% of which were bullocks) from all over rural India, from 81 million in 1971-72 to 77 million in 1991-92 (RUTAG, 2015) -- an outcome of the "Green Revolution" initiated in the late 1960s, whose major benefits resulted in the 1970s and early1980s in substantial increase of agricultural production and in income of farmers (Parayil, 1992), encouraged by commercial banks to purchase tractors (RUTAG 2015), which "plowed the road out of poverty" (ICE, 2000) and replaced bulls, those having been nourished and prospered in Banni grassland included. Thus, while the decreasing flow of forage provision service of Banni ecosystem was driven by local environmental drivers of change, salinity and Pj surges, the reduced use of this service was driven by exogenous indirect technological, economic, social and cultural drivers of ecosystem change.

Note that even though the mainstay of most Banni herders was breeding and trading bulls, this cattle-based livelihood depended on foraging by both cows and bulls. Traditionally, the cow's prime function was milk production for nourishing male offspring, while the function of the female offspring was to produce bulls. Herders, therefore, were hesitant to sell milk, concerned that once started, milk sales might escalate to where the amount of milk left after sales would not be enough for nourishing the calves and would lead to degradation of breed quality (Trivedi, 1965).

However, almost simultaneously with the decline in demand for bulls in India, the demand for cow milk soared, and Banni people, responding to the decline of bullock-based grassland-dependent livelihood, adopted a cow milk-based grassland-dependent livelihood. This happened when India's "White Revolution", another country-wide scale economic and social driver, initiated by a Gujarati milk marketing cooperative by developing a nation-wide milk grid connecting rural producers with urban consumers of milk (Kurien, 2007), arrived at Banni in the mid 1980s. Then Banni cattle owners were included in the regional and national grid, and thereafter, this persuaded the herders, reluctant for centuries to sell milk, to initiate cow milk sales, such that by the 1990s milk delivery from Banni to a dairy became prominent.

This cultural-economic shift from a bull to milk economy brought about a shift in the Banni cattle/buffalo ratio, with the proportion of buffalo increasing from 15% to 45% to 64% in 1960, 1982 and 1992, respectively, and the cattle/buffalo ratio changing from 1/1.4 in 1977 to 1/2.5 in 2007 (Vijay Kumar et al., 2011). This trend, concomitant with those of increasing salinity level and $Pj$ spread, occurred with cows of the Khunni buffalo (Banni breed too, Fig. 3) surpassing the Kankerj cattle cow, and milk superseding traction bulls as the major product of Banni grassland supporting the Banni people's livelihood. This was due to the relative advantage of buffalo over cattle expressed in the buffalo/cattle ratios of milk yield (7-12:1.7 litres/day), milk quality (6.5%:3.5% fat content), and annual income ( $268:$ 154) (Geevanet al. 2003).

While Banni benefited from the White Revolution with its flourishing milk trade in the 1980s and 1990s, circumstances in rural India's Green Revolution changed, bringing a gradual restoration of the country's bull trade, including Banni's. This unanticipated outcome of the Green Revolution became evident when, during the 1990s, India's rural population growth apparently reached a tipping point, and, as a result, land holdings decreased to sizes inappropriate for tractor ploughing. This, and the overlong use of tractors increased soil compactness, diesel pollution and soil biodiversity loss that, together, reduced soil fertility (RUTAG, 2015). These initiated a gradual decline of tractor sales as of 1999 (ICE, 2000) and a bull trade recovery, such that in 2012, 13,000 to 15,000 Banni bullocks were sold (Bharwada and Mahajan, 2012).

Thus, with milk production peaking at 50,000-80,000 litters/day in 2015 (from 46,000 in 2008), and the revival of the bull trade, Banni's annual per capita income (based on the 2001 human population of 16,783) from milk and animal sales combined rose to $518-$791. Nevertheless, Banni's grassland-dependent livestock (cattle and buffalo)-based livelihood could not remain sustainable at the turn of the ${20}^{\text{th}}$ century. Its 7-kg/animal/day grass consumption (Bhimaya and Ahuja, 1969) of Banni's 57,898 adult cattle units (ACU) in 2007 (an annual requirement of 192,994 tons), created a yearly deficit of 18,815 tons (GUIDE 2004). This consumption, defined as overgrazing -- leading to loss of vegetation cover, soil erosion, salinization and further expansion of $Pj$ at the expense of grass covered lands – represents a classic land degradation/desertification scenario.

Added to grassland overgrazing, driven by the forage provisioning service having to support not only bullock but also milk sales, was the differential use of $P_j$ pods and seeds (see 3c) by the two livestock species. As buffalo domination increased, so did $P_j$ seed dispersal and rate of spread, and with this, Banni's rate of transformation from a grassland to woodland ecosystem.

This is because the alien $Pj$ unexpectedly generated an additional provisioning service of Banni's ecosystem - $Pj$ pods provision. The pods ripen during Banni's dry season, when the grass provisioning service declines. Consumption of $Pj$ pods is seriously detrimental to cattle's health, productivity and survival (Tabosa et al., 2003; Kainaat and Jafri, 2015; Azeez and Chandra, 2009), whereas buffalo are resistant to the adverse health effects that cattle suffer. Thus, buffalo, rather than cattle, were the major livestock users of this service, and so $Pj$ spread was further intensified. Furthermore, herders prefer buffalo over cattle not only because of their health resilience and better milk yield and quality, but also because the buffalo's principal diet during the dry period consisted of the Suaeda herbaceous plants that grow in high-salinity sites, which are believed to contribute to the buffalo's milk yield and fat percentage, but which cattle avoid.

Multiple factors advanced $Pj$ 's invasion in Banni: the plant's adaptation for seed germination and spread (see 2c); vicissitudes driven by the Green and White Revolutions in the grassland ecosystem people's livelihoods; and changes in herd composition (see above) and number (increasing from 25,555 in 1977 to 60,417 in 2007, goats and sheep included (Vijay Kumar et al., 2011). The defecation rate of these animals, estimated at about 800,000 dung/day in different locations, drove an increase in $Pj$ seed dispersal. Its spread was moderated by a herders-imposed restriction on the foraging range to village boundaries, given their drinking-water dependence. Although this restriction was relaxed in the early 1990's, in order to increase water availability for livestock outside the village limits, village residents punctured water pipelines constructed by the authorities for connecting Banni villages with a water delivery system. Thus, development policy and cultural drivers further increased the livestock foraging range, and hence, the continued spread of $P_j$.

These effects of increasing negative grazing impact, combined with $P_j$ seed dispersal and spread, led to a rapid transformation of Banni's ecosystem from a grassland to a $P_j$ shrubland/woodland forest ecosystem (officially, Banni has a "forest" status). These, together with the loss of grassland provisioning services that supported both the milk and bullock sale livelihoods, constitute a loss of the biological economic productivity of Banni's land, exemplifying a continuous state of land degradation.

Thus, Banni residents have three options for maintaining a sustainable ecosystem-dependent livelihood and restoring their degraded lands: 1) attempt to block the current trends by fighting the alien invasive tree and restoring the grassland ecosystem's forage provisioning service; 2) explore the prevailing processes for identifying potential services of a $Pj$ forest ecosystem that can generate novel benefits that would support new livelihoods; or 3) search for a balance between these two scenarios by managing Banni as a novel grassland/woodland mosaic ecosystem, with a richer diversity of ecosystem services that support a diverse set of acceptable and sustainable livelihoods, thereby forging an innovative restoration of Banni's land-degraded ecosystem.

V. RESTORATION - TRANSFORMATIVE CHANGE RESPONSE

a) Pj

The mounting damage and projected risk to Banni's traditional grassland-dependent livelihoods driven by $Pj$ invasion brought Banni people to pressure the authorities responsible for introducing $Pj$ to consider totally eradicating it from Banni (see IV 2 "option 1"). However, given the accumulated experiences from the Americas, Africa, Australia and other areas, it was clear that this option would be expensive, and, in fact, unattainable; i.e., only some forms of control – not complete eradication – would be feasible (Pasiecznik et al., 2001). Thus, total eradication was impracticable; however, neither was the alternative of fully abandoning the traditional livelihoods an accepted solution.

This dead-end situation encouraged the emergence of a creative outlet: transforming $Pj$ from the agent of Banni's ecosystem degradation to that of its ecosystem restoration, a solution paradoxically supported by $Pj$ 's invasion adaptations (see III 3). To this was added another $Pj$ attribute of potential demand – the plant's woody biomass, which when manufactured into charcoal, is an excellent firewood in high demand; it burns evenly, does not spark or smoke excessively, and has a high calorific value (4,200 Kcal/kg) – attributes shared with the tree's green branches following a short sun-drying period.

The potential demand for $Pj$ 's biomass materialized when it became a source of domestic fuel for most of the rural households in India; in the last decade of the 20 ${}^{th}$ century, $Pj$ wood supplied about 70% of the domestic fuel of Indian drylands (2.5 tons/year/household). Prior to $Pj$ 's invasion, Banni's people used wood of the land's few indigenous trees as source of domestic fuel. Later, triggered by the increased demand for domestic energy in the Kachchh region around Banni, $Pj$ not only fully supported Banni people's domestic needs, but also gradually endowed them with a new livelihood, based on a novel Banni ecosystem service – a $Pj$ biomass provisioning service, whose benefit is firewood, manufactured into charcoal.

For example, after five years of harvesting all of the $Pj$ trees covering 380 Km ${}^2$ of Banni, yielding an estimated 5,000 tons of charcoal (based on the 2010-2011 fiscal year's data), $Pj$ 's cover of Banni decreased from 31% in 2009 to 16.4% in 2014. Areas from which this wood biomass was removed were simultaneously re-covered by grass, rather than by new $Pj$ plants (Fig.9). This process constitutes a trade-off between the forage provisioning service of the grassland ecosystem with its milk and bulls' sale benefits, with a $Pj$ biomass provisioning service and its charcoal sale benefit. Should this pace of $Pj$ extraction continue, then the dense $Pj$ remaining in 2014 would be converted to charcoal within the next five years – an apparently non-sustainable scenario requiring a change in Banni grassland management.

Based on Banni's dynamic mosaic of low-salinity land plots, tolerated by the indigenous grass vegetation, and others of high salinity, to which $Pj$ is resistant (see iii c. and IV a.), a management plan of (1) removing $Pj$ from the low-salinity plots, thus maintaining the flow of forage provisioning service of these plots, and (2), simultaneously applying a $Pj$ -sustainable harvest regime in Banni's high-salinity plots, could alter $Pj$ 's function as driver of land degradation/desertification, to one of a driver of land restoration and beyond. This would diversify the bundle of Banni's ecosystem services (grass forage plus $Pj$ biomass), adding a new livelihood (charcoal) to the traditional ones (milk, bulls), rather than replacing them.

Thus, an alien invasive species that was a driver of an ecosystem service loss could also function as a biodiversity component underpinning the sustainable provision of a novel ecosystem service, once the demand for this new service emerges. Regarding the sustainability of this new $Pj$ biomass provision, achieved – again paradoxically — by the plant's invasion adaptations of its salinity tolerance and deep-penetrating vertical root system. Together, these enable $Pj$ to suck the highly brackish ground water of the Banni grassland and invest it in fast and vigorous post-cutting re-sprouting. Although cutting at ground level of $Pj$ in an agroforestry system in Sudan slowed the re-sprout process, and their water use efficiency increased after pruning (El Fadl, 1997) in Banni, $Pj$ trees cut at root level regenerated their removed biomass within 3-5 years (GUIDE, 2004, Harsh et al., 1996). Regarding demand – $Pj$ harvesting, responding to an increasing charcoal demand, was practiced during the 1980s though compromised by Banni's State Forest status making this practice illegal, until in 1988, when Gujarat State Forest Development Corporation (GSFDC) allowed selling Banni charcoal, but only to Kachchh residents.

b) Transformative change - livelihood diversification

Before exploring the Banni grassland management option (see V a), the economic benefit of a total $Pj$ eradication as against maintaining the prevailing non-sustainable scenario (Fig.9) had been evaluated in 2016 by The Energy and Resources Institute (TERI) and also reported (Mathur & Sharma, 2018). Based on 23 years of data (1992-2015) this study presents the earnings of the pastoral (milk and livestock sales) and from $Pj$ charcoal, up to the coming year 2030, using system dynamics simulations modelling of two Banni management options – maintaining current (i.e., 2015) policies, versus total $Pj$ eradication.

The variables feeding the model included: spread rate of $P_j$; sizes of livestock herds ( $P_j$ spread multipliers); life table variables of cattle and buffalo; land area under $P_j$ and under palatable grass covers; effect of rainfall on grass productivity; external fodder input, pastoral outmigration; net income from livestock, milk, charcoal and from stress sales. The feedback loops and interactions among these variables governed the model dynamics and the results of its simulations.

These results indicate that maintaining current policies ("business as usual") will leads Banni grassland to severe fodder scarcity driven by grasscover contraction, while total $Pj$ eradication would restore Banni's full grass cover and, by 2030, increase 2015's earnings by 62 per cent. Furthermore, a five-year delay in eradicating $Pj$ would cause a 28 per cent loss of earnings. The study also estimates the cost of totally eradicating $Pj$ during 2016 to 2030, at INR 6,000/ha ${}^5$, would be recovered by income from the restored grass cover, given that the prevailing ban on Banni $Pj$ removal is lifted. In conclusion, attending the limitations of their research, Mathur and Sharma (2018) call for further research into the Banni grassland dynamics and development of policy tools for its sustainable management.

Furthermore, even though Mathur and Sharma (2018) elegantly and persuasively present $Pj$ eradication for recovering the traditional Banni livelihood, considering $Pj$ 's resilience and spread all over Kachchh (i.e. also all around Banni), keeping Banni clear of $Pj$ is apparently not implementable. Yet, maintaining Banni's 1992-2015 management program is unacceptable. Therefore, given the determination of the Maldharis to maintain their tradition and culture on the one hand, and the current gradual spread of the charcoal livelihood in Banni on the other hand, a management for achieving sustainable coexistence of pastoral and charcoal livelihoods based on grass and $Pj$ biomass, respectively, is to be considered.

Indeed, motivated by Banni grassland degradation process and the increasing demand for grass in Banni and beyond, GSFD initiated in 2010 a 20-year Banni Working Plan (BWP), some components of which specifically address the grassland-woodland (i.e.

grass- $Pj$ ) interactions in Banni. The plan does not call for total eradication of $Pj$, but rather for annual harvesting of its sustainable yield from specific areas ( $\sim$ 413.7 Km ${}^2$, 16% of Banni) of its dense cover ( $Pj$ Working Circle areas), and uprooting it from areas ( $\sim$ 673.2 Km ${}^2$, 26% of Banni) in which it is sparse, but the grass cover is dense (Grassland Working Circle areas). $Pj$ plants removed from both areas are to be used for charcoal manufacturing, and, with implementation of two BWP practices, both areas combined would also enable sustainable livestock grass foraging. These practices are: 1) patches of good grass covering $\ge$ 5 ha within the $Pj$ Working Circle would be uprooted at a frequency for preventing their regrowth, thus reducing competition and securing their forage provision flow; and 2) all $Pj$ plants in the Grassland Working Circle area also would be uprooted, at the same frequency, to attain grass regrowth in significant areas of Banni. In the $Pj$ Working Circle areas, where grass is scarce or absent, $Pj$ would be removed rotationally in a regime that would secure a sustainable annual yield for supporting Banni's charcoal economy and emerging tradition.

To accomplish all of this, given that it takes 5 years for $Pj$ regrowth to reach a size of economic value (Gurumurti et al., 1984; Kanzaria & Varshney, 1993; Tewari et al., 2001), a plot from which mature $Pj$ plants have been removed can be re-harvested starting from the $6^{\text{th}}$ year after its initial removal. Therefore, the $Pj$ Working Circle has to be divided into units ('felling series'), each subdivided into 6 subunits (the 5 years of regrowth plus the year of first removal), of 50 ha each ('coupes'); i.e., 138 ([41,370] / [50 ha x 6]) felling series', that jointly making the 'Pj Working Circle'. Thereafter, in each of the five years following the year of removal of one sub-unit ('coupe'), i.e., a $Pj$ cover of 50 ha, will be removed from one of the 'coupes'in each of the 'felling series'. Thus, a cumulative area of 6,895 ha (50 ha x 138 felling units) would generate an annual sustainable yield of $Pj$ economic biomass, an annual contribution of the 'Pj Working Circle' of Banni.

In addition, while adopting the BWP approach -conserving the Banni livestock livelihood by $Pj$ 's sustainable harvesting that allows rotational regrowth of grass, thus supporting both livestock and charcoal livelihoods, GUIDE (Gujarat Institute of Desert Ecology) set to assess the potential of Banni ecosystem and people to achieve that livelihoods' coexistence. This by developing an algorithm (Algorithm 1) for calculating the size of Banni population that could be supported both sustainably (the size of 'block', a term parallel to BWP's 'coupe') and decently (the annual income) / (capita [charcoal worker]) by the current Banni land-cover.

Algorithm 1 first sets the constant values of: Pj density and biomass/ha; Pj biomass-charcoal conversion rate; man-days required for harvesting Pj ha and manufacturing it to kg of charcoal; number of days/year available for harvest and charcoal working; and number of $P_j$ regrowth years.

The algorithm then attends the variables, first those required for calculating/projecting the harvested block size: Banni population size and number of charcoal households; number of persons/household; number of charcoal workers; and the cumulative size of dense $Pj$ area available for harvesting. It then calculates/projects expected income (also compared with India's minimum monthly wage per capita), using the following variables: market price of charcoal/ton; $Pj$ regrowth time (5 years), which together with the year of a block's first harvest, 6 blocks that are to be harvested at 5-year cycles for obtaining an annual sustainable yield. The algorithm thus enables projecting either the number of charcoal households that the area size for $Pj$ harvesting could support, or the size of $Pj$ harvested area that the available charcoal households can harvest.

Algorithm 1 presented here addresses a Banni scenario of 23,000 Banni population in 2011, where 20% of its household (percentage in Banni populations of years 2011-2017, Manjunatha, et al., 2019) include ~766 charcoal households, including 2,300 charcoal makers. This manpower can jointly harvest and manufacture charcoal from six blocks sized 2,760 ha each in a 240-day period of five successive years, (i.e., area of 16,560 ha, 21% of the ~80,000 ha of dense Pj land cover in year 2009). Given the sale cost of charcoal ton (5,000 INR in 2013), the annual value of charcoal harvested of that area (207 million INR), and the number of working days/years, the monthly income per worker (11,250 INR) is higher than the minimum monthly wage in India (9,750 in year 2020).

Algorithm 1. Calculating Banni charcoal-based livelihood's income from sustainably harvesting $Pj$ land cover biomass

Variables for determining 'block' size

Variables for projecting income

Annual income/1 Charcoal Maker (CM) – (207,000,000 INR)/(2,300) =90,000 INR/CM/year Month income/1 CM – (90,000 INR) / 8 month) = 11,250 INR/CM/month = 148 $USinyear2020MinimumIndiamonthwage-9,750INR/CM/month=130$US in year 2019 {"algorithm_caption":[],"algorithm_content":[{"type":"equation_inline","content":"^{1}"},{"type":"text","content":" rounding 45% of 2250 trees/ha (Gavali et al., 2009); (Panday et al., 2012)\n "},{"type":"equation_inline","content":"^{2}"},{"type":"text","content":" mean of 50-80 kg (not including roots) (Sood & Bhatia, 1993); (Khan et al., 1986)\n "},{"type":"equation_inline","content":"^{3}"},{"type":"text","content":" mean of 17-29% (Hall et al., 1982); (Pandey et al., 2012)\n "},{"type":"equation_inline","content":"^{4}"},{"type":"text","content":" 125 days of the year, July to October monsoon period not suitable for charcoal work\n "},{"type":"equation_inline","content":"^{5}"},{"type":"text","content":" Gurmurti, et al., 1984; Saxena et al., 1993; Tewari et al., 2001\n "},{"type":"equation_inline","content":"^{6}"},{"type":"text","content":" rounded 23,065 persons\n "},{"type":"equation_inline","content":"^{7}"},{"type":"text","content":" years 2014-2017, (Manjunatha et al., 2018)\n "},{"type":"equation_inline","content":"^{8}"},{"type":"text","content":" unit area for annual Pj harvesting\n "},{"type":"equation_inline","content":"^{9}"},{"type":"text","content":" year 2009 (GUIDE, 2010)\n "},{"type":"equation_inline","content":"^{10}"},{"type":"text","content":" Indian Rupees\n "},{"type":"equation_inline","content":"^{11}"},{"type":"text","content":" Yadama, 2013\n\nFor maintaining this satisfying scenario projected by Algorithm 1, Banni's demographic trend (for example, 8,724 to 23,065 people from 1961 to 2011), needs to be considered too. For that, projecting the dimensions of Pj land cover that would be sustainably harvested, or – projecting the Banni population size that could be sustainably supported by Banni Pj land cover, need to be assessed."}]}

For example (see Algorithm 2), the scenario of harvesting only 21% of Banni dense Pj area (in year 2009) decently supporting 766 charcoal-based Banni households, suggests that harvesting 13,333 ha (80,000 ha of dense Pj/6 blocks), the maximal Pj block size, would require 11,112 charcoal makers from 3,704 households, whose primary livelihood is charcoal-based. This makes a Banni population of 22,224 persons, a size close to its actual population size in year 2011 (~23,000 people). This projected population size is ~5 times bigger than the 4,600 persons, 20% of the Banni charcoal-based households of years 2014-2017 (given they would have implemented the sustainable harvest scheme spelled out in Algorithm 2).

Furthermore, the results of Algorithm 2 suggest that $Pj$ alone can sustainably support populations of similar sizes, given that the 2009 $Pj$ land cover size is maintained. Nevertheless, these charcoal-based projections are not a scenario of the livestock livelihood's demise. On the contrary, $Pj$ rotational harvest would allow regrowth of grass whose accessibility to the livestock increases, as their foraging movement is no longer blocked by $Pj$ thicket, at least for a period of 2-3 years following a $Pj$ harvest. Also, sustainable yields of $Pj$ charcoal and grass fodder would enable Banni population increase to sizes beyond that of 2011.

Thus, the transformative change scenario would comprise: charcoal workers and livestock herders become beneficiaries of sustainable flows of two different provisioning services of the Banni ecosystems; these Banni grassland ecosystem services are underpinned by two different plant biodiversity components – the indigenous grass species and a tree/shrub species, an invasive alien overgoing a naturalization process. However, for completing the change, field data on grass regrowth, foraging facilitation and rotational grass yield, still requires further research.

c) Past scenarios and the road forward

The previous chapters briefly touched on the Banni grassland conduct for generations until the 1960s $Pj$ introduction, then addressed the undesirable repercussions of this introduction, followed by responses whose realization would require a transformative change, i.e., doing things differently—not just a little more or less of something we're already doing. To fully achieve such a change requires more than bridging the knowledge gaps identified by Mathur and Sharma (2018) and by us (see Algortihm1). Still to be addressed are the environmental and anthropogenic drivers of change revealed throughout the 1980s, 1990s, and the 2000s, which are visually presented here (Fig. 10) by the meandering trajectory of Banni $Pj$ 's charcoal extraction, functioning as an indicator of Banni's dynamic ecosystem and its users' interactions.

Up to the early 1980s, the Banni people sporadically cut $Pj$ not only for domestic use but also for manufacturing charcoal and selling it to traders. Because this harvesting was illegal, its dimensions were not recorded. An additional incentive for $Pj$ cutting emerged when the authorities recognized that, instead of protecting Banni grass land cover from exogenous salt transport (functioning as shelter belt, see iii b), and thus reducing land degradation, $Pj$ negatively impacted this land cover, functioning as a driver of land degradation. Then, as of the early 1980s, a governmental "Banni Development Agency" (BDA), established in the late 1960s, initiated a "grassland development" project for restoring the grassland ecosystem. The BDA contracted private parties to cut down $Pj$ in Banni, not only for encouraging grass protection, but mostly for charcoal manufacturing and selling. This triggered a surge of charcoal tonnage based on cut $Pj$ biomass, legally removed from Banni, that peaked to 300,000 tons extracted in 1988 (Fig. 10).

$Pj$ cutting surge continued until 1995, but this 10-year period experienced climatic oscillations, with drought years ( of 1975-2016 Banni mean annual rainfall), each followed by an extremely high rainfall year ( of mean). These climate-driven oscillations were reflected in $Pj$ availability and charcoal production, and also caused grass reduction that resulted in income loss to Banni people. The latter triggered a $Pj$ -cutting policy change when the Revenue Department -- the government authority responsible for Banni, with its social-economic leaning, requested that the Forest Department, which had outlawed $Pj$ cutting in accord with its role in protecting national forests, will reverse the ban and legalize $Pj$ cutting. This ban lift enabled roadside clearance of $Pj$ cover, a "drought employment" for needy Banni people, which covered their drought years' economic losses, and improved travel safety as well. The rainfall variation and policy changes are expressed by the 1986-1995 oscillations in charcoal production: a range of 150-450 charcoal tons removed in 1987 and 1991, respectively (Fig. 10).

These were followed by restoration of the Forest Department's ban in 1995. Lasting until 2004, it marked the period of lowest charcoal production ( $\sim$ 20,000 ton/year, Fig. 10). But in 2004, the Revenue Department lifted the $Pj$ cutting ban, triggering a ruthless $Pj$ removal, expressed in charcoal manufacturing and income peaks (Fig 10). However, this was soon followed by a steep decline as of 2007/8, when the Forest Department reinstituted the ban in response to the emergence of a charcoal mafia, which was cutting indigenous Banni trees as well as $Pj$, and also to the emerging charcoal cartel's failure to control the charcoal industry. This restored ban was expressed in low charcoal tonnage (25-59 ton/year), followed by declining charcoal market values and income from 2008 to 2016. This regulation to mitigate both illegal and legal $Pj$ cutting was implemented in 2010, with the establishment of the Forest Department's Banni Forest Division; from that time to 2016, 24-25 tons of charcoal were produced annually (Fig.10).

In conclusion, the charcoal trends shown in Fig. 10 depended not only on the grassland ecosystem's natural processes and their responses to the environmental drivers of salinity and rainfall, but also on government policy variations. These are rooted in the strict (though not updated) legal status of Banni as a "national forest", but one in which regulations may be relaxed when it is determined that circumstances affecting the well-being of the locals and the practices for managing their ecosystem require intervention. Moreover, the rather low but relatively stable rate of charcoal production and its support of Banni people's livelihood throughout the second decade of the ${21}^{\text{st}}$ century has been maintained not only by regulation and enforcement, but by its role in a novel emerging socio-economic trend that likely will moderate the Banni people's dependency on its grassland ecosystem provisioning services and their plant products (grass forage for milk and $Pj$ biomass for charcoal). It is based on materialization of potential services of the grassland ecosystem never used before – the Banni ecosystem cultural services.

Banni is kind of a terrestrial island away from India's bustling cities - a mid-desert productive greenery of diverse land covers (trees, shrubs and herbs, water bodies), inhabited by a visually and vocally tangible biodiversity of birds (262 species, Koladiya et al., 2016), mammals, bees and butterflies (among them 'charismatic'), and by indigenous human communities of rich culture and skills adapted to life in this unique ecosystem. These attributes of the Banni ecosystem enrich it by adding to its already used ecosystem service bundle, a potential cultural service that would materialize once the demand for it has been created. Globalization and other cultural-socioeconomic processes have apparently already created this demand, seen in the appearance of bhunga huts(a traditional mud structure) and tourism in Banni in the first decades of the 21st century. Added to this are cultural attractions, such as the "white desert" festival and, with support of the Gujarat Tourist Department, a trend in homestay/ecotourism, currently with over 200 tent houses in Banni. This suggests that the income generated by the Banni grassland's cultural services even at present would slightly lessen the people's critical dependence on its provisioning services (fodder and $Pj$ biomass).

Returning to the road forward, it should be noted that since its initiation in 2010, the 20-year BWP has been engaged in circles not addressed in this paper, e.g., Improvement (plantations, etc.) and Wildlife (habitats) circles. More relevant, the BWP's Grassland Working Circle has, already advanced grassland restoration throughout its first decade, work that will likely continue, together with work in the $Pj$ working circle, as planned, until 2030. Then, being a longstanding document BWP will function, with some additional amendments as per the scenario in 2030.

Furthermore, all BWP working plan areas are to be overlapped by unique working circle – the Joint Forest Management (JFM) Working Circle, constituted to provide active participation of local people in the management of Banni. The JFM involves local villagers working with forest officials, guided by a joint committee of the two groups. Presently the JFM is active only in the Grassland Circle, but it is expected to work for a mutually agreed upon program of conservation, restoration and management, along with regulating the collection of minor forest products and $P_j$ removal (where needed), as well as charcoal manufacture. With this, it is likely that once the JFM has become more active, the sustainable harvest algorithms (see V b) would be promoted for sustained implementation. Thereby, a diversity of Banni sustainable livelihoods can be maintained, based on the grassland's bundle of ecosystem services, underpinned by an invasive alien that has become a naturalized biodiversity component of the Banni grassland ecosystem.

VI. BANNI LAND DEGRADATION AND RESTORATION

a) Pj and land degradation

This paper examines approximately 60 years of the Banni Grassland, an ecosystem at the edge of the Asian Thar desert, which hosts and supports an indigenous human community, modulated by environmental- and human-generated vicissitudes. We spotlighted the role of a shrub/tree invasive alien, Prosopis juliflora, in the dynamics of this unique and intriguing human-environment system. For this, we reviewed literature addressing rangelands and their people, how they responded to challenges of dryland climatic conditions, land degradation, and alien species' invasions, and the efficacy of the responses. Some of the literature focused on Pj invasions mainly in Africa and India, and others looked specifically at Banni.

Most reports point to knowledge gaps that require bridging through further research. For example: "Research is needed to identify novel solutions to aid the dilemma of management and contentious issues regarding invasive Prosopis globally" (Zachariades et al., 2011); and "...dilemmas and conflicts of interest among stakeholders involved in the management of P. juliflora need to be thoroughly investigated" (Ayanu et al., 2015). For our review to incentivize further studies to bridge the knowledge gaps regarding grasslands whose lands experience degradation, invasion of alien species and restoration, we used an integrated conceptualization of these processes by streamlining them into the

Biodiversity and Ecosystem Services (MEA ${}^6$ ) and Land Degradation and Restoration (IPBES' LDRA ${}^7$ ). This approach is presented here as the Conceptual Framework of Banni Grassland Socio-ecosystem, addressing Banni ecosystem dynamics in three stages, outlined in Figures 11a, 11b, and 11c, respectively, as follows: Prior to Pj introduction; Pj invasion – grass-land degradation; and Pj naturalization – socio-ecological system restoration/rehabilitation.

Adapted IPBES Conceptual Framework (Diaz et al., 2015; IPBES, 2018) - boxes and arrows denote the elements of the Ecosystem and the Human system and their links indicated by numbered arrow are described in the main text.

In Fig. 11a, the boxes in the lower (green) rectangle represent the ecological system landscape ${}^8$ and those in the upper (gray) rectangle represent the human system landscape. Together, they function as a unified socio-ecosystem, with their linkages depicted by an ellipse of arrows starting from the bottom center

Ecosystem box, leading to the top center Human system box (arrows 1,2,3), and back to the Ecosystem Box (arrows 4,5,6). An explanation of the contents of the boxes follows, starting with the Ecosystem Box.

A desert land blessed with a perennial grass cover, Banni is categorized as a grassland ecosystem – an Ecosystem Type common in drylands. Its substantial interacting components, soil and biodiversity– comprise Banni's natural capital, generating a flow of Ecosystem services (arrow 1). This flow is supported by ecosystem supporting functions –primary production and nutrient recycling. These functions create and maintain the biomass of Banni grass species, a component of Banni biodiversity, underpinning provisioning and cultural services delivering a flow of Benefits, thus linking the natural ecosystem to the human system (arrow 2). These benefits supplied by the ecosystem support and strengthen the well-being of Banni people (arrow 3).The major benefit of the grass forage provisioning service is feeding the Banni livestock, thus supporting the "Banni bullock" sale as the mainstay of Banni people prior to $Pj$ 's introduction, to which sale of ghee, milk product of buffalo (mainly) and cattle, locally very popular. These benefits constitute the economy and health components of Banni people's well-being. Furthermore, the Banni pastoral tradition of breeding bulls is embedded in the culture of the Banni herders, a culture based on the grass cover of the Banni ecosystem. In the ecosystem services context, these are the Banni grassland ecosystem's cultural services - non-material benefits that are often expressed in people's sense of place and social cohesion, and contribute to both mental and physical health, all components of human well-being (Human system box).

The flow of the provisioning service is subject to drivers of ecosystem change. These are often initiated by indirect, underlying drivers generated by people and elicited by the state and/or changes in the well-being of people, local and beyond (arrow 4). Shown in the Human drivers of ecosystem change box are the most Banni-relevant - policy, governance and economy, that are indirect, underlying human drivers of change; i.e., the Banni human system affecting its ecosystem (arrow 5). These include the human-dependent sporadic overgrazing in Banni, and interlinked environmental drivers, such as rainfall variations: i.e., severe rainstorms and floods directly cause soil erosion, and droughts increase soil erosion, while rainstorms mitigate the soil salinization driver underpinned by Banni inherent underground saline water, affecting plant biodiversity (red arrow parallel to arrow 6). This driver of ecosystem change endangers the Banni grassland ecosystem's state (Ecosystem box, top), though its state prior to $Pj$ introduction has been categorized as sustainable; i.e., amply maintaining the well-being of Banni people, but always at risk of amelioration due to the soil salinization potential threat.

Returning to the Indirect drivers' box, intangibles, such as institutions, governance and technology are linked (arrow 7) to education and local knowledge systems, collectively designated "Anthropogenic assets", which are instrumental in materializing and making accessible the ecosystem-derived benefits that contribute to the people's well-being, and represent an ecosystem and people co-production of benefits (arrows 8 and 3). For example, the breeding and sale of bullocks depend on veterinary services and local knowledge if this benefit is to contribute to the sustainability of Banni people's well-being. Before leaving Fig. 11a we return to the Ecosystem box and its 'sustainable' state, since natural processes, i.e., seasonal rivers' freshwater flow through Banni, reduce Banni's prevailing inherent salinity, thereby maintaining it as a grassland. But this sustainability is at risk, given that the 'soil salinization' direct driver may become active, and put at risk the striving for human well-being (arrows 3).

With this in mind, we move forward in time, to Fig. 11b – the socio-ecosystem dynamics triggered by $Pj$ 's introduction (see IV b) and subsequent invasion. Here we start with the Human system box, affected by water shortage (in this case, off Banni), prompting (arrow 1) a policy of river damming (see III b), a process that intensifies soil salinization, and thereby transforms an environmental direct driver of soil salinization to a human one (Direct drivers' box of Figures 11a and 11b, respectively). This is amplified by the indirect human drivers of ecosystem change (Indirect drivers Box, arrow 2) affecting the two components of the ecosystem (arrow 3) – soil, whose salinity impacts biodiversity.

The increased salinity and the awareness of its expected negative impact on Banni's human well-being (Human system box) triggered a second policy initiative – the introduction to Banni of an invasive alien, $Pj$ (arrow 4), which functioned as a direct human driver of ecosystem change. Within a short time, the invasion of the alien plant impacted the ecosystem's biodiversity components (arrows 5) by outcompeting the Grass biodiversity component and reducing its land cover in Banni (see III c). The impact on the grass cover that underpins the supply of the provisioning and cultural ecosystem services (arrow 6), conferred the title "degraded" on the ecosystem's State (top of Ecosystem box). The degradation impacted the services' benefits (arrow 7) (which at that time also included milk sale [See IV c]), and thus created potential harm to the well-being of the Banni people (arrows 8), i.e., degradation of the human system in response to the ecosystem degradation. Furthermore, the grass degradation process enabled a gradual $Pj$ grass cover replacement through a process of "ecosystem transformation" from a grassland ecosystem to a $Pj$ woodland/shrubland ecosystem (Arrow 9), which in turn led to a further ecosystem transformation (see Fig. 11c).

Thus, Fig. 11b presents the 'invasive alien species' as a direct driver of degradation processes, often addressed in the context of land degradation – eighth in the list of land degradation drivers and defined as "the state of land which results from the persistent decline or loss in biodiversity and ecosystem functions and services" (IPBES, 2018), or "persistent reduction in the capacity of ecosystems to supply services" (Millennium Ecosystem Assessment, 2005), either through biodiversity loss, or by transformation to a derived ecosystem type" (IPBES, 2018). Hence, given that the degraded title is assigned to state of a grassland ecosystem, this state of Banni grassland can qualify as 'land degradation', and since Banni land is climatically classified as arid dryland, this land degradation process also qualifies as desertification (see IV b).

b) Pj naturalization and land restoration

Here we move forward in time from the $P_j$ invasion process and the grassland degraded state (Fig. 11b), to $P_j$ naturalization and the socio-ecosystem restoration processes (Fig. 11c). Although some confusion persists in the literature regarding 'naturalization' vs 'invasion' in alien and/or invasive plant species' stages, it is proposed that naturalization starts "when barriers to survival and then barriers to regular reproduction are overcome" (Richardson et al., 2000). However, earlier definitions of a naturalized species, e.g., "Introduced from another region but growing, reproducing and maintaining itself in competition with the native vegetation" (Walker, 1989), and "established extensively amongst native vegetation so as to appear native" (Clement & Foster, 1994), fit well with $P_j$ 's status as shown in the Direct drivers of ecosystem change (box of Fig. 11c).

Furthermore, following the linguistic definition of naturalization – "the admittance of a foreigner to the citizenship of a country" (Oxford languages, https://languages.oup.com/), and given that the Banni authorities promoted and the community adopted, the 'control by use' approach (Van Wilgen et al., 2012; Shackleton et al., 2014) rather than the Pj eradication one (see V a & V b), the naturalization stage of Pj can be defined as "admittance of alien invasive species to, and acceptance by, Banni's indigenous socio-ecological system". This, and given that already "in several areas of South India, Prosopis trees and shrubs have become naturalized" (Pasiecznik et al., 2001), Pj has already been naturalized in Banni too. And, it functions as an Indirect driver (arrow 1), converting the Direct driver of degradation to a driver of rehabilitation (arrow 2), a component/state of Restoration that falls short of fully restoring the grassland ecosystem to its pre-degradation state (IPBES, 2018). Thus, the degraded Banni grassland ecosystem, expressed in reduced land areas of grass cover, has not been restored to the pre-invasion land size of grass cover, but has been rehabilitated to function as a mosaic of grassland/woodland/shrubland ecosystem, jointly covering the pre-invasion Banni grassland ecosystem area, now in a rehabilitated state (top of Ecosystem box).

This diversified mosaic of grass patches, dense $Pj$ tree patches and sparse $Pj$ shrub (young $Pj$ plants) patches constitutes a novel grassland/woodland/shrubland ecosystem supplying a diversified bundle of ecosystem services (Arrow 3). Most important for the Banni people's well-being is the provisioning service, now underpinned by two biodiversity components: an herbaceous plant community, many of which are livestock-palatable (see IV b), and labelled here as "grass"; and a biodiversity component comprising the by now naturalized $Pj$ plant –trees and shrubs of all ages and densities. Thus, the novel ecosystem supplies two ecosystem provisioning services – grass and $Pj$ products, livestock forage, and woody biomass, respectively. The benefits of each provisioning service (Arrow 4) include milk, a critical product enabling bulls to be sold as truck bullocks, and milk sale for human consumption; and $Pj$ woody biomass for manufacture and sale of charcoal as an energy source.

Furthermore, the two biodiversity components, grass and $Pj$, underpin additional ecosystem services, cultural and regulating, respectively. Since the pastoral livelihood of the herders is based on the ecosystem's grass cover that nurtured their pastoral tradition which they cherish, it qualifies as one of the nonmaterial benefits from ecosystems, known as 'cultural services'. Grass, in addition to its forage function, is also valued for its green landscape, contrasting with the surrounding desert. This, together with other biodiversity components of Banni grassland, especially the 'charismatic' species (e.g., Houbara, Flamingo, Chinkara [Indian gazelle], Desert fox and others), provides another cultural service, eco-tourism (see V c), which benefits tourists with recreation, relaxation and spiritual fulfilment options, and benefits Banni people with tourist fees, an income added to that of their sales of milk and charcoal products.

In addition, Banni's biodiversity underpins a bundle of regulating services, such as regulation of soil erosion caused by intense rainfall and wind storms, as well as livestock trampling and overgrazing. This erosion is mitigated by the below- and above-ground Banni land cover of grass and of $Pj$ plants, where the share of $Pj$ to the regulating service is significant, due to its below-ground root systems and the above-ground cover of its woody canopy ( $Pj$ was introduced in the 70s to Ethiopia for addressing soil erosion problems, [Shiferaw et al., 2004]). Furthermore, $Pj$ 's role in the soil erosion regulation service made the plant an agent of full restoration of Banni's economic biological productivity (not just rehabilitation). This in specific land areas where this productivity had been lost due to human induced intensification of their soil salinity, i.e., areas of land degradation, what makes them ones whose biological productivity is restorable.

This degradation/restoration process occurs in Banni land areas where inherent high salinity levels have been transformed to an extreme salinity level (see Table 1, in III c), driven by the human driven river damming process. Since most grass species of these land areas are intolerant to that high salinity, these lands have lost their grass cover – the biodiversity underpinning the soil erosion regulation service flow, and they become exposed to soil erosion. Thus, two drivers of ecosystem change – human induced soil salinity followed by soil erosion – jointly underly an instance of 'land degradation' in Banni. However, $Pj$ is resilient to these areas' extreme salinity, so that its cover, even if sparse (see Table 1), would underpin the soil erosion regulation service flow. Thus, once processes of salinity mitigation would become operational, with $Pj$ cover functioning as a mitigator of soil erosion, the economic productivity of the land would be restored; this exemplifies a case of $Pj$ functioning as a driver of full land restoration in some areas of Banni grassland.

In addition, $P_j$ also underpins the global climate change regulation service through the process of carbon sequestration, wherever it covers Banni. Here again its above- and below-ground live and dead biomass, of dimensions that enable engaging and benefiting from the carbon trade, is an as of yet unrealized benefit. Note that for the benefits presented in the Benefits Box that would support human well-being, mainly the Banni economy component, the anthropogenic assets (listed in this box, and more) need to be used (arrow 5 and then 6). Additional regulating services underpinned by $Pj$ (not presented in Fig. 11c) are water flow and purification regulation (Walter & Armstrong, 2014). Finally, the current and future state of Banni's socio-ecological system as presented in Fig. 11c, lies in the hands of the Human system to control its underlying drivers of change (arrow 7), such that the grassland's ecosystem state and its services, as presented in Fig. 11c, will continue to sustain an acceptable level of the Banni people's well-being.

Furthermore, while Figures 11a, b, and c show the dynamics within the network of inter-linkages of the variables and processes of Banni's socio-ecosystem, moving from sustainability, to degradation, and then restoration, Table 3 adds a final dimension to the story; that is, the chronology of events that shaped the trends of processes in the Banni socio-ecosystem (i.e., whether they increased, decreased, or remained stable), and their vicissitudes over six decades, from 1960 to 2015.

Table 3: Chronology of Banni processes' vicissitudes, 1960 – 2015

The table, based on our review data, presents the trends of ten Banni variables and the processes and drivers underlying these changes. One variable (milk sale), emerged only in the 3rd decade, and two others (Pj biomass and charcoal sale) in the 4th decade; in both cases, their growth stabilized after the first decade of their appearance. Soil salinity is the sole variable that remained stable throughout the early 1960s, and then it continually increased throughout the entire review period. Altogether, the trends of these four variables changed twice.

Trends of the interlinked grass cover and forage provision changed 4 times, bullock sales 5 times, and the remaining variables, 3 times each. Finally, in the last review decade (the 2010s), seven of the ten variables remained stable, and three – the ecosystem services' benefits, i.e., sales of milk, charcoal and bullocks, continue contributing to the resilience of Banni's human well-being, even though grass cover has not yet been fully stabilized.

Given all this, it is likely that Land Degradation Neutrality (LDN), i.e., "an ecosystem's state whose functions and services that promote food security, remain stable" (UNCCD, 2015), is achievable, in Banni and in other sites experiencing similar circumstances.

VII. CONCLUSIONS AND RECOMMENDATINS

Our review describes a rolling course of socio-ecological crisis management of a reputable rangeland in Asia, the Banni Grassland in India, which was threatened with land degradation by soil salinization resulting in reduced biological productivity on which this ecosystem users' livelihoods depend. The salinization process was generated by two distinct, but mutually independent drivers, of which authorities were aware. Nevertheless, from the beginning, management focused only on the first driver (windborne salt particles from an adjacent saline desert). This approach not only failed to moderate the strength of the driver, but it heightened the threat to the biodiversity component (livestock's edible grass plants), and thus to the grassland people's livelihood and lifestyle. Meantime, management ignored the more active driver - stoppage of the annual freshwater flow by damming the rivers that open into Banni and leach the soil surface salt deposited through capillary movement of Banni saline groundwater.

The outcome of the actions taken in response to the crisis led us to two conclusions/recommendations: first, do not look for solutions until you have identified and fully recognized the drivers of the crisis; and second, do not use a biodiversity component as a remedy until you have fully studied and understood its natural history and its function and behavior in its habitats and ecosystems. These two conclusions/recommendations are not new and may even seem trivial, but our 60-year review of the Banni grassland ecosystem adds to their support and strength, as based on crisis management of a single grassland ecosystem, for which the chosen remedy was an alien plant, one that is also a notoriously invasive species.

Moreover, Banni's case shows that effective management of an ecosystem requires not only literacy of the used ecosystem and of the species that are candidates for responding to drivers of its degradation, but also an understanding of its human system: both the people living in the ecosystem and using its services, their culture, needs and aspirations; and the authorities – local and external – who have influence and power in matters concerning the ecosystem, its users and their well-being. These are reflected in the third conclusion/recommendation: an effective response to ecosystem degradation requires familiarity with and knowledge of, the ecosystem-human system interactions.

These recommendations are based on knowledge gained from 60 years of trial-and-error wrestling with an invasive alien, widely known to be harmful as an 'ecosystem engineer unleashed' (Ayanu et al., 2015), which managed to reduce the flow of the major provisioning service of the invaded ecosystem. The lessons learned enabled conversion of the invasive alien to a biodiversity component underpinning a novel ecosystem service, thus offsetting the partial loss of the grass provisioning service, and fully restoring the land's biological productivity of economic value, while maintaining the traditional indigenous livelihood, diversifying the ecosystem service bundle, and increasing the users' livelihood resilience.

Thus, our review presents a case of tackling land degradation based on retaining critical levels of natural capital; i.e., lands of tolerable salinity and their grass cover, while basing the ecosystem users' livelihoods on a wider range of ecosystem services. Our review, therefore, supports a vision of sustainable coexistence of grass and $P_j$ in a spatial land mosaic, and diversity of sustainable livelihoods in the Banni grassland. This is in contradiction to science-based recommendations for Banni proposing that ".... land under P. juliflora needs to be cleared" and "... Prosopis removal policy... is a favorable policy option" for Banni (Mathur & Sharma, 2018). Another science-based study, however, proposes to "... strengthen the positive aspects of Prosopis, a naturalized resource that has come to India to stay" (Walter & Armstrong, 2014). Our advice to the bewildered, therefore, is to attend available experiences yet not give up until you've explored other "out of the box" approaches.

Regarding this advice and noting that Banni case of restoration/rehabilitation of an Indian desert grassland ecosystem degraded by an alien species invasion is not unique; further search may reveal similar but not identical cases. Here we present one case of "out of the box" solution, the rehabilitation of an Inner Mongolian grassland ecosystem, Hunshandake Sandland (HS) (King C., 2008). In this case the degradation is caused not by an invasive alien species, but by overgrazing. In both Banni and HS the direct consumers of the grass cover are livestock, mostly cattle, but the loss of forage in Banni was caused by the alien Pj outcompeting the forage plants, while in HS, overgrazing has limited the ability of the land to carry enough animals to sustain the livelihoods of local families.

In neither case has the pre-degradation stage been restored; neither Banni's total land's grass cover, nor HS's grass productivity per land unit. Rather, in both cases the ecosystem has been rehabilitated by human-driven introductions; of an invasive alien plant species in Banni, and of a less destructive livestock species (chicken) in HS. Both Banni and HS benefited from the new species, which yielded diversified income sources for livelihoods; the sale of charcoal in Banni and of chicken eggs in HS. Thus, in both cases, the ecosystems' rehabilitation fully restored and even increased the income from the ecosystems, i.e., the ecosystem-based human well-being of the locals. And, in both cases the deep-rooted attachment of the local herders to the pastoral tradition – i.e., a benefit of the ecosystems' cultural service – has been maintained, such that altogether, full restoration of human well-being achieved in both cases, maybe both through 'out of the box' solutions.

Returning to the Banni case, we suggest that the Conceptual Framework of Banni Grassland Socio-

Ecosystem – the Banni-adapted conceptual framework of biodiversity and ecosystem services (Figs.11a, b, and c), can function as an algorithm for moving from natural capital (the ecosystem's biodiversity) to human well-being, and from degraded land towards its sustainable management. This, together with our two algorithms (see V b) "Calculating Banni charcoal-based livelihood's income from sustainably harvesting $Pj$ land cover biomass", and "Projecting Banni population size sustainably supported by Banni $Pj$ land cover", can function as effective guidelines for implementing a successful transition from non-sustainable to sustainable ecosystem use. For example, in the case of Banni, use of these guidelines saw $Pj$ transformed from an agent of land degradation to an agent of land restoration, as a result of residents' and authorities' informed joint intervention.

Furthermore, these guidelines may be useful not only at the local level, but also at regional and global levels, such as contributing to the global intergovernmental effort for achieving, by 2030, the United Nations Sustainable Development Goals – and specifically its 15.3 target. These call for "restoring degraded land and soil" and striving to "achieve a land degradation-neutral world", when success would be based on comparing the state of functions and human well-being of a focal socio-ecosystem in 2030 to those prevailing in 2015, the last year of our Banni review.

ACKNOWLEDGEMENTS

The authors wish to acknowledge Dr Arun Kumar Roy Mahato for providing references and calculations pertaining to the charcoal data of this review, and Mr Dayesh Parmar for preparing satellite maps and their data analysis, and Mrs Marian Segal, for her language editing of this paper.