The study was carried out in Upper and Lower Casamance, an area in the south of Senegal. These two zones cover an area of approximately 28350 km², or 1/7 of Senegal’s surface area ( Figure 1 ).

The Ziguinchor region, located in southwestern Senegal at 12˚33'N latitude and 16˚16'W longitude, spans 7,339 km², accounting for 3.73% of the country’s territory. It borders The Gambia to the north, Guinea-Bissau to the south, the Kolda and Sédhiou regions to the east, and the Atlantic Ocean to the west. Ziguinchor experiences a tropical climate with distinct dry (November to May) and rainy (June to October) seasons due to the African monsoon, making it one of Senegal’s wettest regions, with annual rainfall between 1300 and 1500 mm. The average annual temperature is around 27˚C, with maximums reaching 37˚C in April and minimums dropping to 15.5˚C in January. The region’s soils are predominantly hydromorphic in valleys and sandy or sandy-clay ferruginous on plateaus, with generally flat terrain [13] .

The Kolda region, situated in southern Senegal at 12˚45'N latitude and 15˚30'W longitude, covers 21,011 km², or 10.68% of the national territory. It borders The Gambia to the north, Guinea and Guinea-Bissau to the south, Tambacounda to the east, and Sédhiou to the west [14] . Kolda has a Sudanian climate with two distinct thermal seasons. The cooler season from July to February sees temperatures range from 20.26˚C to 30.36˚C, while the hotter season from March to June experiences temperatures between 21.26˚C and 35.40˚C. The region receives an average annual rainfall of 1191 mm, with considerable variation. The landscape features sandy-clay sandstones forming plateaus with savannah or open forest, interspersed with valleys containing rice fields and lowland pastures [3] .

The survey of herbaceous species was conducted in July and August in Ziguinchor, located at 12˚78'33"N latitude and 16˚21'66"W longitude. The Braun-Blanquet phytosociological method was employed, which involves compiling a list of species present within a representative and homogeneous sample of the herbaceous cover. Survey sites were selected based on the homogeneity of the herbaceous vegetation to minimize edge effects [15] . Additionally, all recorded herbaceous plants were identified both in the field and later in the laboratory using references such as the Flora of Senegal [16] and Tropical Weeds [17] .

An interview guide was developed in collaboration with the heads of the departmental technical services in Ziguinchor, Kolda, and the Kolda Zootechnical Research Centre (ZRC). Using quota sampling, a survey was conducted among livestock farmers to gather relevant data. A structured questionnaire was administered to farmers raising cattle, sheep, goats, and pigs in the two municipalities during the winter season, aiming to collect information on the quality of fodder for each herd and the grasses most consumed by the animals. In total, sixty farmers were interviewed thirty in Ziguinchor and thirty in Kolda.

Some of these farmers were closely monitored as they led their cattle into grazing areas, such as the bush or surrounding regions, to observe the animals’ preferences for available grasses and to compare these observations with the farmers’ reports. During these monitoring sessions, key activities such as grazing, moving, drinking, resting, and rumination were recorded. The herder also assisted in identifying the herbaceous species most favored by the cattle.

To assess the overall value of the rangelands, two key indicators pastoral value and carrying capacity were employed [18] . Pastoral value, a method for evaluating the forage quality of rangelands [19] - [21] , depends on both the specific contribution of species (SCi) and their specific quality index (SQI). The SQI is determined either by the bromatological values of the plant species found in the analyzed environments or by the scores provided by the farmers, shepherds, herdsmen, and other resource persons interviewed.

They were entered into the Sphinx software and processed using Excel. To assess the diversity and floristic composition of the species in relation to their pastoral value, we sought to determine their cover, frequency and the contribution of each species.

$\text{SFPi}\left(\%\right)=\frac{\text{ni}\times 100}{\text{N}}$

ni: number of points where species (i) is present. N: total number of read points.

$\text{SCi}\left(\%\right)=\frac{\text{SFPi}\times 100}{\sum \text{SFPi}}$

$\text{Cpv}=\frac{1}{3}\sum \text{SCi}\times \text{SQI}$

Cpv = Crude pasture value. SCi = Specific Contribution. SQI = Specific Quality Index.

The majority of respondents (86.7%) in both Ziguinchor and Kolda were aged between 30 and 60 years ( Table 1 ). The remaining age groups, those between 0 and 30 years (5%) and those 60 years and older (8.3%), showed significant differences in representation. The sample was predominantly composed of Peulh (55.7%), Diola (36.7%), Manding (3.3%), Sereres (2.3%), and a small percentage (2%) belonging to other ethnic groups such as Akou, Manjack, and Lebou. The significance of these ethnic groups varied depending on the type of livestock farming practiced.

Agriculture was the primary occupation for 40% of the respondents, closely followed by animal husbandry at 38.3%. Other activities, including trade and various professions (civil servants, mechanics, etc.), accounted for 21.7% of the respondents. The gender distribution among livestock farmers indicated that the majority were men (86.7%), with women making up 13.3%.

In terms of income, the largest proportion of respondents reported earning between 50% - 75% of their monetary income from livestock farming (48.3%), followed by those earning 25% - 50% (43.3%), more than 75% (6.7%), and 0 - 25% (1.7%). Respondents also highlighted the role of livestock in their household livelihoods as very important (48.3%), important (31.7%), medium (18.3%), and low (1.7%).

Membership in farmer organizations was noted among 46.7% of respondents, while 59.3% reported not being affiliated with any such groups. Additionally, 45.7% of the farmers had another significant source of income, whereas 54.3% relied solely on livestock farming.

Livestock farming was the second most important socio-economic activity in Casamance after agriculture. Livestock include cattle (local breeds, Mathia, Gobra), sheep (local breeds, mixed breeds), goats (local breeds, mixed breeds) and pigs (local breeds, Large White, Landrace). The livestock population was large and diverse ( Table 2 ).

Our findings revealed that cattle, sheep, and goat rearing held significant importance among the Peulh, whereas pig farming was predominantly practiced by the Diola and other ethnic groups. Pig rearing, in particular, involved a notable proportion of the Diola community, especially among women. Field observations indicated that the highest animal densities were found in the Kolda community.

Across both zones, traditional extensive and semi-intensive livestock production systems were predominant, accounting for 56.7% of practices ( Figure 2 ). These systems were mainly adopted by the Peulhs and Diolas. However, the Diola community is currently facing several challenges, including the insecurity of pastoral land tenure and the scarcity of fodder. Additionally, 43.3% of the farms were engaged in a combination of extensive, semi-intensive, and intensive animal husbandry.

Among the pastoralists surveyed, 8% of the Fulani did not have cattle in their herds. For those who did rear cattle, 98% of the herds consisted of 0 to 60 head, with only 2% having at least 80 head. Regarding sheep, 20% of the farmers, including both Peulh and Diola, had no sheep at all. Additionally, 72.7% of the flocks had between 0 and 20 sheep, while 27.3% had fewer than 20 head. Goats were present in many of the herds, with 63% of the herds containing between 0 and 20 goats and 36% having fewer than 20 head.

Pig ownership was predominantly among the Diola and other ethnic groups in both regions. However, some respondents did not rear pigs. Of those who did, 80% had between 0 and 15 pigs, while only 20% had more than 20 pigs. It was observed that some herds included one, two, or three categories of animals, often featuring a mix of local breeds, crossbreeds, or animals produced through artificial insemination ( Figure 3 ).

The herds of the primary ethnic groups surveyed were acquired through various means: inheritance (40.3%), purchase (50.4%), gift (7.3%), and loan (2%) ( Figure 4 ). Inheritance was the predominant method for the Peulh, whereas the Diola and other ethnic groups primarily acquired their herds through purchase. During the dry season, herds were typically grazed either within the village or town or in the surrounding forest. In contrast, during the winter, most cattle were moved to more distant grazing areas or open forests.

This revision clarifies the methods of herd acquisition and seasonal grazing patterns, ensuring the information is concise and easy to understand.

Monitoring showed that animals were engaged in activities such as grazing, watering and ruminating. A total of 67% of farmers used this practice, with 64.3% using natural grazing only and 35.7% using supplemental feeding.

The survey identified a total of 15 herbaceous forage species palatable to ruminants, categorized into 7 families ( Table 3 ). The Poaceae family was the most abundant. Among the 15 species, 9 were notably palatable, with frequencies exceeding 10%. These included: A. gayanus, highly cited (equal to 50%); Aeschynomene indica, Eleusine indica, Cynodon dactylon, Digitaria horizontalis, Panicum maximum, Pennisetum pedicellatum, Dactyloctenium aegyptium and Aeschynomene americana. These species were cited with moderate to high frequency (ranging from 50% to 10%). The remaining 6 species—Sida rhombifolia, Ipomoea tiliacea, Andropogon hispidum, Hyptis suaveolens, Mariscus cylindristachyus, and Paspalum scrobiculatum—were cited less frequently.

NB: D = Diola; P = Peulh.

Among the species observed in the rangelands, three species—Panicum maximum, Ipomoea tiliacea, and Aeschynomene americana—were not included in the list of inventoried species.

The grassland flora survey revealed 95 species, distributed across 68 genera and 22 botanical families ( Table 4 ). The most common genera were Ipomoea (7 species), Cyperus (4 species), and Eragrostis (3 species). Key contributors to the herbaceous cover were Momordica charantia, Cenchrus biflorus, and Oplismenus undulatifolius.

The total herbaceous ground cover was 58.3%. Twelve species exceeded an average cover of 10%, including: Momordica charantia (30%), Cenchrus biflorus (26%), Oplismenus undulatifolius (19%), Setaria barbata (17.33%), Dinebra retroflexa (17%), Dactyloctenium aegyptium (16.49%), Cyperus squarrosus (15%), Ipomoea heterotricha (15%), Leucas aspera (15%), Digitaria horizontalis (14.55%), Setaria pumila (11.18%), Dioscorea bulbifera (11.13%).

Species like Momordica charantia (6.92), Zornia glochidiata (5.54), Spermacoce sp (4.85), and Aeschynomene indica (4.15) demonstrated the highest specific contributions, although their relative values were generally low ( Table 4 ).

The dietary spectrum analysis ( Figure 5 ) revealed that Lpv species were the most significant with a Specific Contribution Index (SCI) of 52.31% and a Specific Quality Index (SQI) of 1. Mpv species followed with a SCI of 37.05% and an SQI of 2. Together, these two categories accounted for 89.36% of the total index values. In contrast, Npv plants were completely neglected (0% SCI) and Gpv plants had a low overall presence with a SCI of 10% and an SQI of 3.

Surveys revealed that 64.3% of ruminants were fed exclusively on natural pasture during the rainy season, while 35.7% received supplementary fodder for sick, lactating and pregnant animals. Among farmers who practiced fodder collection and conservation, crop residues comprised the majority of stored fodder, averaging 85.4% ( Figure 6 ). Natural straw (5.8%), natural tops (6.8%), and cultivated tops (12.8%) were less common compared to cereal stalks (72.6%), which were of fair quality. Fodder plants (1.1%) were grown in small quantities and constituted marginal stocks.

Farmers cited various reasons for not cultivating fodder crops, including seed issues, lack of knowledge on cultivation techniques, financial constraints, limited time and insufficient land. The quality of stored fodder varied: cereal stalks were fair, legume tops were good and herbaceous fodder was excellent. Notably, Alysicarpus ovalifolius and Dactyloctenium aegyptium hay, cowpea and peanut tops were well preserved. In contrast, Zornia glochidiata tops, which were often cut after losing their leaves, had average quality compared to other preserved legume species.

Forage collection periods varied by category. Natural fodder was collected year-round, with peaks during the main and post-harvest seasons, while crop residues were gathered and conserved immediately after harvest. During the dry season, natural fodder stocks mainly included stems and leaves of herbaceous species.