The importance of being reliable – Local ecological knowledge and management of forage plants in a dryland pastoral system (Morocco)

1

This is the post-print version (post-refereeing, i.e. the final version of the article that the

authors have seen before it was published) of the following article: Linstädter, A.,

Kemmerling, B. (equally contributing), Baumann, G., Kirscht, H., 2013. The importance of

being reliable – Local ecological knowledge and management of forage plants in a dryland

pastoral system (Morocco), Journal of Arid Environments 95: 30-40, which has been

published in final form at http://dx.doi.org/10.1016/j.jaridenv.2013.03.008

The importance of being reliable - valuation and management of forage

plants in a dryland pastoral system

Anja Linstädtera*, Birgit Kemmerling

b, Gisela Baumann

a,c, Holger Kirscht

d,e

a Range Ecology and Range Management Group

Botanical Institute

University of Cologne

Zülpicher Straße 47b

D-50674 Cologne, Germany

b

Institute of Oriental Studies

University of Leipzig

Schillerstraße 6

D-04109 Leipzig, Germany

E-mail address: [email protected]

c Present address:

Botanic Garden and Botanical Museum Berlin-Dahlem

Freie Universität Berlin

Königin-Luise-Straße 6-8

D-14195 Berlin, Germany

E-mail address: [email protected]

d

Institute of Social and Cultural Anthropology

University of Cologne

Albertus-Magnus-Platz

D-50923 Cologne, Germany

e Present address:

International Institute of Tropical Agriculture Cameroon

1 Main Road IRAD Nkolbisson

P.O. Box 2008 (Messa)

Yaoundé, Cameroon

E-mail address: [email protected]

* Corresponding author

Dr. Anja Linstädter

Tel. +49-(0)221-470 7905

Fax +49-(0)221-470 7908

E-mail address: [email protected]

A. Linstädter and B. Kemmerling contributed equally to this study.

2

Abstract

In dryland rangelands with their high environmental variability, local knowledge of forage

plants is essential for management decisions. Ecological apparency hypothesis (EAH) predicts

plants’ availability and visibility to be important criteria for local valuation. However, EAH

has mainly been tested in low-variability systems. We ask whether EAH is valid for forage

plants in drylands; which other local criteria exist; and how criteria are connected to

management decisions.

In a Moroccan pastoral system, we applied a novel ethnobotanical method by calculating the

Cognitive Salience Index (CSI) both for plants’ valuation (CSIantro) and availability (CSIeco).

To evaluate explicit criteria, we correlated palatability and nutritive value to CSIanthro.

ANCOVAs related CSIanthro to EAH criteria (CSIeco and lifetime) and to plant occurrence on

pasture types. We found EAH criteria to better predict CSIantro than explicit criteria. Apparent

plants from semi-arid pastures were more valued than those from arid pastures (HSD; p <

0.05). We introduce reliability as one criterion into EAH to explain this, and demonstrate how

pastoralists adjust management decisions to resource reliability. Linking resource valuation to

management decisions can thus improve our understanding of resilience mechanisms. Our

study also confirms the validity of EAH for forage species and dryland environments.

Keywords: Arid and semi-arid rangelands; Cognitive salience index; Ecological apparency

theory; Local ecological knowledge; Pastoralism

3

1. Introduction

In arid and semi-arid environments, a sustainable land management requires a profound local

ecological knowledge (LEK) of natural resources and their dynamics (McAllister et al., 2006).

For this reason, LEK and its integration in scientific knowledge has received increasing

attention in dryland science (e.g. Fernandez-Gimenez, 2000; Thomas et al., 2007; Whitfield

and Reed, 2012). In the case of pastoralism, LEK is particularly required on the spatio-

temporal variability of forage resources (McAllister et al., 2006; Thomas and Twyman, 2004).

Based on their herding experiences, cognitive abilities, and world views, pastoralists gain an

extensive body of LEK on forage resources. This includes the environmental conditions in

which a plant grows, its palatability depending on livestock species, its phenology and life

history, and its availability on local pastures (Fernandez-Gimenez, 2000; Thomas and

Twyman, 2004). All parameters may be relevant for the perception of a species’ value as

forage plant. However, quality judgements are often not explicit, but part of implicit or even

tacit knowledge (Fazey et al., 2006), which makes them empirically difficult to extract

(Wesuls and Lang, 2010). To explain the valuation and use of plant resources in local

communities, the ecological apparency hypothesis (EAH) has been among the most promising

approaches (Lucena et al., 2007; Lucena et al., 2012). It was initially proposed for herbivore

behaviour and has been later adapted to ethnobotanical research questions (Phillips and

Gentry, 1993a, b). The hypothesis claims that apparent plants i.e., the most visible, most

dominant, and most frequent plants will have a higher cultural importance than less apparent

plants, not because they are necessarily more useful, but simply because they are more

available or visible to human communities (Lucena et al., 2007). However, EAH has mainly

been tested in tropical systems with a comparatively low spatial and temporal variability, such

as humid forests (Albuquerque and Lucena, 2005; Phillips and Gentry, 1993a, b), and dry

forests (Lucena et al., 2007; Lucena et al., 2012). It thus remains unclear which implications

environmental variability has for the validity of EAH.

The perception of natural resources and their dynamics is the basis for pastoral management

decisions (Fernandez-Gimenez, 2000; Steele and Shackleton, 2010). In a given situation, a

particular action is chosen from a set of strategies. Herd mobility is the most important

pastoral strategy, as it enables herders to cope with effects of rainfall variability on the natural

resource base (Butt, 2010; Dwyer and Istomin, 2008). Pastoral-nomadic range management

may have negative effects on the sustainability of natural resources (Bollig and Schulte, 1999)

but it often has positive effects (Müller et al., 2007b) and increases the resilience of the

pastoral system to external shocks such as meteorological droughts (Dougill et al., 2010;

Linstädter et al., 2010). In this paper we focus on a dryland pastoral system along a steep

gradient of environmental variability to test the validity of EAH for forage plants in highly

variable environments. We apply a novel, quantitative methodology to analyse the local

valuation of forage plants and to ask whether it is based on explicit local knowledge or

ecological criteria. Explicit LEK is analysed as the ‘palatability for goats’, ‘nutritive value for

goats’, ‘palatability for sheep’ and ‘nutritive value for sheep’. As ecological criteria related to

EAH we test a plant’s availability on local pastures, its life form and its lifetime. We then

relate management decisions to the local valuation of pasture quality, i.e. to the availability

and quality of forage resources on pasture types. Besides mobility decisions, we evaluate

alternative strategies of herd management which may be connected to the local valuation of

forage resources. Our key questions are:

- Does EAH explain the local valuation of forage plants in a highly variable environment?

- Do pastoralists have other explicit or non-explicit criteria for the valuation of forage plants?

- How are these criteria connected to herd management decisions?

4

2. Methods

2.1. Regional context and study area

Data were collected on Morocco’s southern slopes of the High Atlas Mountains among the

pastoral-nomadic group of the Ait Toumert. The Ait Toumert who settle in the Mgoun district

speak Tachelhit, one of the three Berber languages of Morocco. So far, little research has been

conducted in that area and no recent official census data were available on household level.

Through interviews with local leaders, 29 households of the Ait Toumert have been identified,

some of them sedentary and not applying transhumance anymore, others sedentary, but

sending their herds with other herdsmen, while some are still pastoral-nomadic. We restricted

our sample to pastoral-nomadic households. These use sheep and goats for livestock

production, while donkeys, mules and dromedaries are used for transportation (Akasbi et al.,

2012). The normative transhumance cycle (see Linstädter et al., 2010) runs along a steep

altitudinal gradient, from the mountainous summer pastures at 3,000 m asl (used from mid-

May to September) to the lowland pastures at 1,300 m asl used during the winter months from

November to February. In October and again from March to mid-May nomads pass the

intermediate pastures in between (Fig. 1).

Fig. 1. Grazing area of the Ait Toumert (High Atlas, Morocco), with four pasture types arranged along

a gradient of increasing aridity (from the summer pasture in the northern highlands to the far winter

pasture in the Basin of Ouarzazate). Pasture types were identified by means of expert interviews and

mental maps. Study sites are Trab Labied (TRB) for far winter pastures, Taoujgalt (TAO) for near

winter pastures, Imeskar (IMS) for intermediate pastures, and Tizi n’Tounza (TZT) for summer

pastures.

Summer and intermediate pastures are exclusively used by Ait Toumert households and are

subject to strong institutional arrangements. Summer pastures are temporarily protected by

means of the agdal institution. The agdal is a communal pasture; opening and closing dates

are fixed by the community of users (Genin and Simenel, 2011). Access to the summer

pasture is closed from mid-March to mid-May, and generally restricted to members of the

5

tribal fraction with access rights. The comparatively vast winter pastures have an open-access

tenure regime and are shared with at least two neighboring fractions. The schedule of

transhumance may deviate from this normative cycle, depending on rainfall, herd dynamics,

and on socio-cultural factors (Akasbi et al., 2012).

Table 1. Location, climate, and rangeland characteristics of the four study sites representing

the Ait Toumert pasture types. The map codes of study sites refer to Fig. 1. MAP is the mean

annual precipitation. Tmin is the mean daily minimum temperature of the coldest month, Tmax is

the mean daily maximum temperature of the warmest month. The UNEP aridity index AIU is

the ratio of MAP to annual potential evapotranspiration (Middelton and Thomas, 1997).

Bioclimatic unit follows Oldeland et al. (2008), the vegetation type is taken from the

vegetation map by Finckh and Poete (2008). ANPP is the above-ground primary production,

measured (like standing crop) for the growth period 2007/08 within temporal exclosures.

Grazing offtake is the percentage of ANPP consumed by livestock during that period; it

indicates recent grazing pressure. For details of standing crop and ANPP measurements refer

to Linstädter and Baumann (2013).

Ait Toumert Pasture/ Far Winter/ Near Winter/ Intermediate/ Summer/

Study site (map code) TRB TAO AMS TZT

Location

Latitude 31°10’ N 31°23’ N 31°29’ N 31°34’ N

Longitude 6°34’ W 6°19’ W 6°14’ W 6°17’ W

Altitude (m asl) 1380 1870 2250 2960

Abiotic site conditions

Bioclimatic unit Arid cool Arid cold Semiarid cold Subhumid cold

MAP (mm a-1

)* 124 170 285 363

Tmin coldest month (°C)** 9.2 3.2 3.7 -2.2

Tmax hottest month (°C)** 30.7 25.5 23.4 16.5

UNEP aridity index AIU* 0.06 0.12 0.16 0.22

Rangeland characteristics

Vegetation type Saharan

Hammada

semidesert

Ibero-Maurita-

nian Artemisia

steppe

Juniperus

woodsteppe

Oromediterra-

nean shrubland

Total standing crop (kg ha-1

) 331 889 565 8786

Woody perennials (% biomass) 59.8 80.5 81.8 98.0

Herbaceous peren. (% biomass) 1.6 10.0 7.8 0.6

ANPP (kg ha-1

yr-1

) 244 560 224 419

Grazing offtake (% of ANPP) 57.5 91.4 71.5 56.3

Abiotic site conditions: * = calculated for the years 2000 – 2006 (01.01. to 31.12.), ** =

calculated for the hydrological years 2001/02 – 2007/08.

Sheep and goats share the same pasture areas and are jointly herded. Herdsmen do not specify

their herd management for livestock species; they merely adjust their micro-mobility on local

pastures to the proportion of goats within a herd to account for the different feeding needs of

the two ruminant species. While herds dominated by goats easily access rocks and scarps, and

eat more woody plants such as cushion-like xerophytes and dwarf shrubs, herds dominated by

sheep are not able to exploit a pasture in that way. In consequence herdsmen have to be more

mobile to meet the feeding needs of sheep.

The altitudinal gradient from the central High Atlas Mountains to the basin of Ouarzazate is

also a steep gradient of environmental variability from a semi-arid to an arid environment

6

(Table 1). Similar to other altitudinal gradients of decreasing aridity (e.g. Zemmrich et al.,

2010), typical vegetation belts are distinguishable. In the High Atlas Mountains, they are

roughly congruent to the altitudinal levels of the four pasture types. The mountainous summer

pastures from 3,000 to 2,500 m asl are covered by an Oromediterranean shrubland dominated

by persistent, cushion-forming shrubs. Intermediate pastures running from 2,500 to 2,000 m

asl are mainly covered by a Juniperus woodsteppe. The near winter pastures cover the

altitudinal belt from 2,000 to 1,500 m asl. This is the zone of an Artemisia steppe, dominated

by dwarf shrubs and perennial grasses. The Saharan Hammada semidesert of the far winter

pastures (from 1,500 to 1,000 m asl) is dominated by annual species, interspersed with shrubs

such as Hammada scoparia (Benabid and Fennane, 1994; Finckh and Poete, 2008; Linstädter

and Baumann, 2013).

2.2. Implicit and tacit criteria for the valuation of forage plants

Within the context of testing EAH, the use-value index has been proposed as a quantitative

measure for the relative importance of species to a human population (Phillips and Gentry,

1993a, b) and frequently been applied (Albuquerque and Lucena, 2005; de Almeida et al.,

2005; Lawrence et al., 2005; Lucena et al., 2007; Lucena et al., 2012). However, this index

was not applicable in our case as we were not interested in multiple uses of plants, but in

species’ valuation as forage plants. Moreover, we aimed to record quality criteria including

those which are implicit or tacit. For this purpose, tools of Cultural Domain Analysis such as

scoring and ranking methods have been proposed (Bernard, 1994; Wesuls and Lang, 2010).

We chose the free-list technique (Borgatti, 1999) to gain quantitative information on the

valuation of forage species. While free-lists have mostly been used to elicit the items within a

cultural domain, they can also be analysed in terms of an item’s salience (Borgatti, 1999).

Salience analysis assumes that informants will arrange items in terms of individual saliency,

and that tacit and/or implicit criteria will determine saliency (valuation). The free-list method

has also been recommended for ethnobotanical purposes, specifically to find out the most

culturally salient plants of a particular use category (Quinlan, 2005).

We separately questioned 17 informants from 15 households to name all forage plants they

knew to be eaten by sheep and goats. Informants, normally the male head of the household,

represented the complete sample of pastoral-nomadic households that could be met on

summer and intermediate pastures during fieldwork. In three cases, women were interviewed

and in one case the son of the household-head. All households owned the animals they were

herding, and all informants were or had been engaged in tending herds. The average age of

interviewees could not be identified, as indications of birth were based on estimations due to

the lack of birth certificates. It can be assumed that the majority of informants were between

35 and 55, with one informant significantly younger and one significantly older. Some of the

informants had gone to Quran School, but none had a formal education. Those who could

afford it had started to send some of their children to school. Two translators from the region,

familiar with the socio-cultural context and with interviewing techniques assisted in

interviews which were conducted in Tachelhit. To maximize output in free-listing tasks, we

used two interviewing techniques suggested by Brewer (2002), i.e. a nonspecific prompting,

and reading back the list of free-listed items.

Although more than 20 interviews had been planned, only 17 could be realized as we were

constrained by the great difficulty to track nomadic households by foot in their pasture areas.

The sample thus misses the recommended minimum size of 20 to 30 interviews (see Sutrop,

2001 and references therein). However, this recommendation refers to the situation when only

term frequency is evaluated to establish a cultural domain. If the mean position of an item is

7

additionally considered (which we did with Sutrop’s cognitive salience index; see below),

smaller samples are sufficient (see Sutrop 2001:272).

We aimed to minimize potential bias by conducting interviews at different times and in

different localities. Most interviews were conducted by visiting the pastoral-nomads at their

current housing or on pastures (intermediate or summer pastures) in September and October

2007. At this time of the year pastoral-nomads were moving from the summer pastures to the

lower intermediate pastures, two of the smaller pasture areas where informants could be

tracked within walking distance. Upland pastures of the Ait Toumert fraction were hardly

accessible by car. Conducting interviews at two different pasture areas also served to

minimize bias by the fact that informants could tend to name species first growing on these

pastures. 45% of the informants were met on summer pastures while 55% had already moved

to intermediate pastures. As informants did not stay on winter pastures during the field study,

interviews could not be conducted there. It would also have been hardly possible to track Ait

Toumert households during winter months, when herders mix up with those of other fractions

throughout the vast areas of winter and intermediate pastures.

The results of individual interviews were later (in April 2008) supplemented and triangulated

by expert interviews with herdsmen from different households. The expert interviews were

carried out as group interviews with male herders, using a general interview guide approach.

In addition the responsible person (sheikh) for formally opening the agdal was questioned in

an individual interview. The interviews took place at the weekly market near Ait Khalifa.

During that time, households stayed on intermediate pastures which allowed an easy and

regular access to the market and the gathering of the assembly of Ait Toumert representatives

to discuss the opening of the agdal.

2.3. Explicit criteria for the valuation of forage plants

To extract explicit LEK on the valuation of forage plants, we conducted semi-structured

interviews with Ait Toumert herdsmen directly after free-list tasks. We asked what

characterizes a species as valuable forage plant for sheep and goats. Interviews rendered two

criteria roughly similar to the scientific concepts of palatability and nutritive value.

Palatability relates to plant characteristics or conditions stimulating a selective response by

animals, with palatable species being preferentially eaten (Baumont et al., 2000). The

nutritive value refers to whether a plant species will promote rapid weight gain and animal

health (Komwihangilo et al., 2001).

In a later step, we recorded local estimations of palatability and nutritive value for 62

abundant species representing different growth forms on local pasture types. This was done

during a group interview with three knowledgeable herdsmen in October 2007. Informants

were asked to classify species into ‘refused’ (0), ‘little or occasionally eaten’ (1), ‘regularly

eaten’ (2) or ‘highly favoured’ (3). Informants also classified species’ nutritive value as ‘bad’

(0), ‘medium’ (1) or ‘good’ (2). These categories resulted from the preceding interviews on

explicit criteria for the valuation of forage plants (see above). Preference ranking was done

separately for sheep and goats.

2.4. Vegetation sampling

We recorded vegetation parameters reflecting plants’ occurrence on pasture types, and their

ecological apparency (visibility and availability). Based on vegetation studies (Benabid and

Fennane, 1994; Linstädter and Baumann, 2013) and own observations, plants were assigned

to the pasture type where they mainly occurred. To record the availability components

abundance and frequency, we selected four sites representing the four vegetation belts along

the altitudinal gradient (Fig. 1). Sites were located close to weather stations established by the

IMPETUS research project in 2000. Hence, we were able to record specific climatic

8

conditions of study sites (Table 1). Even though sites were mostly positioned at the border of

or adjacent to the Ait Toumert pasture area, they were representative for the vegetation

composition the used pasture types, as the altitudinal belts of pasture types are roughly

congruent to vegetation zones. Species cover values were visually estimated on 25 m²

vegetation plots. We sampled 24 plots in each of the four pasture types (total: 96 plots),

keeping a minimum distance of 20 m between them. On each site, data were collected at the

end of the local growth period (i.e. for the time of peak biomass), which was between March

2008 (winter pastures) and June 2008 (summer pastures).

Visibility was recorded as life form and life history (lifetime). We distinguished three life form

classes, i.e. therophytes (herbaceous annuals), hemicryptophytes (herbaceous perennials) and

woody species, subsuming chamaephytes, nano-phanerophytes and phanerophytes (Table 2).

For lifetime, we ranked plant growth forms (see Table A.2 for data) according to life duration,

from short-lived annual plants (1) over perennial forbs (2), perennial grasses (3) and short-

lived woody plants (dwarf shrubs; 4) to shrubs (5) and persistent woody plants, i.e. cushion-

like xerophytes and trees (6).

2.5. Management decisions

We conducted semi-structured interviews to evaluate how the local valuation of forage

resources was linked to management decisions. To consider environmental variability and its

influence on resource availability, 16 of the 17 informants interviewed with the free-list

technique were asked to classify forage availability in past years as being very good, good,

average, bad, or very bad. We did not suggest any standard with respect to an ‘average’ year,

and interviewees were asked to go back in their memory as far as possible.

To match herders’ reports on available forage to precipitation records, we used data for two

subregions of the broader study area with similar rainfall characteristics. The classification

was based on the Standardized Precipitation Index (SPI) of hydrological years (August of the

preceding year to July of the following year), calculated for all available Moroccan weather

stations at a given time (Born et al., 2008).

As a local classification of years may, besides the availability of forage resources, reflect

other aspects such as the individual economic situation, qualitative information was used to

distinguish ecological from economic reasons. Informants were asked to recall individual

mobility decisions and alternative strategies of herd management in four hydrological years.

As neither precipitation records nor informants’ reports indicated a year with ‘very good’

resources in the close past (Table 5), we concentrated on years named as either ‘good’,

‘average’, ‘bad’ or ‘very bad’ for resource availability. With respect to management

decisions, we focused on the winter months (October to March), because this is the time of

resource scarcity in the annual cycle (Linstädter and Baumann, 2013) when local knowledge

on quality and availability of forage resources should be most important for management

decisions.

2.6. Data analysis

2.6.1. Match of local and scientific plant names

For free-listed plants only known by their vernacular, informants were asked to collect

samples for a taxonomic identification. Missing vernacular names were obtained in interviews

with local informants. Additionally, ethnobotanical publications on Moroccan plant species

were consulted (e.g. Iav, 2002; Trabut, 2006). The reconciliation of vernacular and taxonomic

9

names is prone to confusion (Szabó, 2004). We distinguished five cases (see Table A.1): (i)

plants with a clear match, (ii) synonymy, i.e. more than one vernacular for one scientific

name, (iii) homonymy, i.e. one vernacular for several scientific names, (iv) more than one

vernacular for several scientific names, and (v) no scientific information for a vernacular. In

case of synonymy, the most salient vernacular was used for further analysis. When homonyms

occurred, ground cover values of all homonymous species on a certain plot were accumulated.

If several vernaculars were related to several scientific names, plant groups were formed.

Here, 33 vernaculars and 84 scientific names were summarized into eleven plant groups. If no

scientific information was available for a vernacular, it was excluded from further analysis.

We also omitted all items that were mentioned only by one informant to improve the

robustness of the list task parameter ‘term frequency’ (see Sutrop 2001). The final

anthropological data set consisted of 79 identifiable items (out of 109 occurring on two or

more free-list, see Table A.1). In the ecological data set, 103 out of 133 items were identified

(see Table A.2) and occurred more than once on plots.

2.6.2. Match of anthropological and ecological data sets

We used a novel ethnobotanical method to match free-list (anthropological) data and

ecological data. Both data sets were parameterized in the same way with a weighted rank

index, i.e. the Cognitive Salience Index (CSI) proposed by Sutrop (2001) for free-lists. CSI

calculates the salience (S) of an item based on its frequency (F) and mean position (mP) on

free-lists, and the number (N) of informants, where S = F/(N mP). Values range between zero

and one. Plants with high CSI values are listed most commonly by informants, and are

recalled more immediately than other plants (Thompson and Juan, 2006). To calculate CSI

values for the ecological data set (CSIeco), we set CSI parameters defined for free-lists as

analogous to parameters of vegetation plots:

1a. Individual free-lists of N interview partners on forage species to

1b. Plant species lists of N vegetation plots (including non-forage species)

2a. Mean rank of a species in each free-list (mP) to

2b. Mean rank of a species on each plot (mP) (abundance, measured as ground cover)

3a. Species’ frequency in all free-lists (F>2) to

3b. Species’ frequency on all plots (F>2).

In analogy to CSIanthro which quantifies a forage plant’s salience (local valuation), its mean

abundance (mP) on vegetation plots together with its frequency (F) determines its availability

(CSIeco). We used the statistical software ANTHROPAC 4.0 (Borgatti, 1996) for all CSI

calculations (see Table A.2 for CSI values).

2.6.3. Statistical analysis of explicit LEK

We treated CSI data and the index of lifetime as continuous. To improve normality and

homogeneity of variances, CSIanthro and CSIeco data were log-transformed, and data on plant

lifetime were z-transformed before analysis. As ‘palatability for sheep’, ‘nutritive value for

sheep’, ‘palatability for goats’, and ‘nutritive value for goats’ were on an ordinal scale, we

used Spearman’s rank correlation coefficient to correlate them to each other and to CSIanthro.

2.6.4. Statistical analysis of implicit LEK

For the intersecting data set of plants with non-zero CSI values in both the anthropological

and the ecological data set, we used general linear models (ANCOVAs) to test the effects of

the factors ‘pasture type’ and ‘life form’ and the linear predictors ‘lifetime’ and ‘CSIeco’ and

their interactions on the response variable CSIanthro (Table 2). We first fitted a full model,

10

including all variables and two-way interactions. Model simplification involved a stepwise

backward selection based on the Akaike Information Criterion (AIC). This goodness-of-fit

statistic considers both fit and model complexity; it penalizes models for each explanatory

variable, i.e. for greater complexity (Johnson and Omland, 2004). Analyses were performed

using the software R 2.13.1 (R Development Core Team 2011).

Table 2. Parameters tested as predictors for the local valuation (Cognitive Salience Index,

CSIanthro) of forage plants, and their connection to criteria of Ecological Apparency

Hypothesis (EAH).

Parameter Acronym Description Range EAH criterion

Ecological

performance

CSIeco Weighted rank based on

a plant’s abundance and

frequency on plots

0.01-0.27 (potential: 0-1) Availability

Life form LF Raunkiaer life form,

chamaephytes and

phanerophytes grouped

as ‘woody’

Therophytes (annuals)/

hemicryptophytes (perennial

forbs and grasses)/ woody

plants (shrubs and trees)

Visibility

Plant

lifetime

Lifetime Index of plant life

duration

1 (annuals) -6 (trees) Visibility

Occurrence

on pasture

types

Pasture Pasture type where

plant mainly occurs

Summer pasture (SU)/

intermediate pasture (TR)/

near winter pasture (NW)/

far winter pasture (FW)

Not related to

EAH

3. Results

3.1. Explicit LEK on forage plants

The four explicit local criteria for plant quality (‘palatability for sheep’, ‘nutritive value for

sheep’, ‘palatability for goats’, and ‘nutritive value for goats’) were not significantly

correlated to CSIanthro values (Table 3). In contrast, all local criteria were correlated, with the

highest correlation between a plant’s nutritive value for sheep and its nutritive value for goats.

Pastoral nomads did thus not clearly distinguish between a plant’s forage value for sheep and

goats.

Table 3. Correlation (Spearman's rank correlation coefficient) between explicit local criteria

for forage plant quality (palatability and nutritive value for sheep and goats) and the local

valuation of 62 fodder plants, quantified with the Cognitive Salience Index (CSIanthro).

Significant correlations (p < 0.01) are marked with an asterisk.

CSIanthro Palatability

for sheep

Nutritive value

for sheep

Palatability

for goats

Nutritive

value for goats

Palatability for sheep 0.054 1.000

Nutritive value for

sheep

0.282 0.667* 1.000

Palatability for goats 0.167 0.536* 0.457* 1.000

Nutritive value for

goats

0.151 0.454* 0.681* 0.642* 1.000

11

3.2. Implicit and tacit LEK on forage plants

From the two complementary data sets on forage plants, we identified an intersecting set of 49

forage plants that had two non-zero CSI values. In the minimal adequate model (Table 4),

CSIanthro values of these plants depended on three out of four explanatory variables included in

the full model.

Table 4. Minimal adequate model for the local valuation of forage plants (quantified with the

Cognitive Salience Index, CSIanthro) as a function of the predictor variables ‘pasture’ (presence

of plants on local pasture types), ‘lifetime’, ‘CSIeco’ (plants’ ecological performance

quantified with the Cognitive Salience Index) and ‘life form’ (not included in the minimal

adequate model). Effect sizes are given as Eta-squared ( ). For details on predictor variables

refer to Table 2.

Parameter SS Df MS F P

Pasture 8.29 3 2.76 5.34 0.003 0.17

Lifetime 7.99 1 7.99 15.48 < 0.001 0.16

CSIeco 6.70 1 6.70 12.99 < 0.001 0.14

Pasture : Lifetime 4.86 3 1.62 3.14 0.035 0.10

Residuals 21.16 41 0.52

The occurrence of plants on local pasture types explained the highest proportion of variance

in the dataset (16.9%). Plants mainly growing on summer pastures had significantly higher

CSIanthro values (Fig. 3) than plants occurring on all other pasture types (Tukey’s HSD; p <

0.05). Plant lifetime had a significant positive effect on the local valuation of forage plants

(Fig. 2, right), and explained 16.3% of the variance in the dataset. Similarly, availability also

had a strong positive effect on CSIanthro (Fig. 2, left) but explained slightly less variance

(13.7%).

Fig. 2. Relationship between the local valuation of forage plants (local weighted ranks,

CSIanthro) and the two EAH criteria ‘availability’ (left; quantified via ecological weighted

ranks, CSIeco) and ‘visibility’ (right; quantified as an index of plant lifetime). Solid black lines

are linear regressions for all data points (left: f(x) = 0.471x, p < 0.001, R² = 0.232; right: f(x)

= 0.349x, p < 0.001, R² = 0.302), coloured lines indicate separate regressions for pasture types

if coefficients significantly differed from zero (p < 0.05; R² = 0.546 for near winter pasture,

0.447 for summer pasture, 0.295 for intermediate pasture).

12

We found that the factor life form and all interactions except the two-way interaction between

pasture type and lifetime did not influence local valuation of forage plants (p > 0.05). The

significant interaction (pasture type: lifetime) confirms that regression coefficients differed

between pasture types (p < 0.05; see Fig. 2 for linear regressions between CSIanthro values and

lifetime indices specified for pasture types).

Fig. 3. Local valuation (log CSIanthro) of forage plants on the four pasture types of the Ait

Toumert, arranged (from left to right) along a gradient of decreasing environmental

variability. FW = far winter pasture, NW = near winter pasture, TR = intermediate pasture,

SU = summer pasture. Boxes show medians and 25th

to 75th

percentiles, whiskers stand for the

non-outlier ranges of the data. Different superscripts (a, b) indicate significant differences

(Tukey’s HSD; p < 0.05).

3.3. Forage availability and management decisions

While some herdsmen were able to recall the quality of years back to the mid 1970ies, most

herdsmen only remembered the past four to six years. Thus, the frequency of mentions

declined for more distant years, with 51 % of all statements referring to the past two years.

Although herdsmen referred to the entire hydrological year, it became clear throughout the

interviews that the winter months with the erratic precipitation patterns on winter pastures

were crucial for the classification of years. The local perception of available resources showed

little variation for the past four years (2003/04 to 2006/07) (coefficients of variation < 0.4; see

Table 5) but variation increased for the years 2002/03 and 2001/02. Reports for prior years

generally concerned ‘very bad’ years (results not shown). We matched precipitation records

from two subregions of the Ait Toumert pasture area to herder reports (Table 5) and found

that reports tended to be quite accurate, although very sensitive to low precipitation in any of

the two subregions. Individual mobility decisions during winter considerably deviated from

the normative transhumance cycle (Fig. 4), which would be to move to near winter pastures in

November, and then – in years with good resource availability – to far winter pastures in

December. In years remembered as average and bad, households often remained on

intermediate pastures and on far winter pastures. The near winter pastures were preferably

used in good, but also in very bad (drought) years. Households exhibited different patterns of

intra-seasonal movements on pastures. Whereas transhumant mobility between different

pastures was practised in average and bad years, a constrained mobility occurred during good

and very bad years: households tended to stay on one pasture type. Decisions for a

constrained mobility were often combined with new strategies of herd management, such as

feed supplementation, and truck transport, or a mass selling of animals.

13

Fig. 4. Macro-mobility of Ait Toumert households (n = 16) during the winter months of four

years which were recalled as either ‘good’, ‘average’, ‘bad’ or ‘very bad’ with respect to

available forage resources (A-D; Very Bad Year = drought year; see Table 5).

Table 5

Herders’ assessment of forage availability on local pastures for the six most recent growth

periods (LEK-based classification), compared to a record-based classification of precipitation

which uses the Standardized Precipitation Index (SPI) of hydrological years for two

subregions of the study area with similar rainfall characteristics (see Born et al., 2008), i.e.

SOA = the semi-arid region south of the Atlas Mountains, and ATL = the Atlantic region,

including high elevations of the Atlas Mountains. SPI classes follow the categories proposed

by McKee et al. (1993), with an additional indication of above-average (+) and below-average

(-) precipitation for near-normal years. LEK classes are based on the median of individual

assessments (‘very bad’ years scoring as 0, and ‘very good’ years as 4). Coefficients of

variation (CV; the ratio of the standard deviation to the mean) are given as a measure of the

relative variability of assessments.

Year LEK-based classification Record-based classification

N Mean CV Median LEK class N

(SOA)

SPI class SOA N

(ATL)

SPI class ATL

2006/07

13 2.46 0.356 3.00 Good

10 Near normal (+) 10 Near normal (+)

2005/06

10 1.40 0.369 1.00 Bad

10 Near normal (-) 12 Severely dry

2004/05

5 2.20 0.203 2.00 Bad

10 Moderately wet 13 Near normal (+)

2003/04

4 1.50 0.385 1.50 Average to bad

11 Near normal (-) 13 Near normal (+)

2002/03

4 1.75 0.719 2.00 Average

11 Near normal (-) 16 Near normal (+)

2001/02 4 0.75 2.000 0.00 Very bad 11 Severely dry 15 Moderately dry

14

During the winter months, constrained mobility was a common alternative to transhumance in

all types of years (Fig. 5). Some informants stated that they did not move to far winter

pastures in spite of good forage availability because they had already bought supplementary

feed for the winter. Feed supplementation had been introduced by the Moroccan government

during a drought in the 1980s. Today pastoralists have adopted it as a regular strategy for the

unpredictable winter months: About two-thirds of informants stated that they provided

supplementary feed during winter months, irrespective of the availability of natural resources.

As households could buy supplementary feed on a market close to the intermediate or near

winter pastures, they tended to stay there in average, bad or very bad years. The truck

transport constitutes another new strategy to react more flexibly in drought years (Fig. 5).

While the majority of households remained on the near winter pastures of the Ait Toumert

territory, some households with access to alternative income sources could afford to rent a

truck, and transported their herds to far-off pastures. The remaining households often sold

their animals.

Range management decisions were clearly based on the set of strategies available to

households, which has been constantly changing and included new strategies introduced

during the last decades. While all households moved to summer pastures, mobility decisions

during winter months were taken based on the perception of forage availability (see above)

and on a household’s resource portfolio such as income, manpower and social networks. For

example, the access of winter pastures required manpower and transport animals, while a

constraint transhumance during winter months required feed supplementation which needed

monetary resources (see Breuer (2007) for more detailed information on monetary,

incorporated and socially institutionalized resources in pastoral-nomadic groups of the High

Atlas Mountains).

Fig. 5. Proportion of Ait Toumert households (n = 16) which had applied alternative strategies

during the winter months of four years which were recalled as either ‘good’, ‘average’, ‘bad’

or ‘very bad’ with respect to available forage resources. As households could apply several

strategies during the same year, proportions may sum up to > 100%.

4. Discussion

4.1. Explicit and implicit LEK on forage plants

We tested explicit criteria for the local valuation of plants as forage, as well as non-explicit

criteria reflecting a plant’s ecological apparency and its occurrence on pasture types. Our

results show that the explicit local criteria ‘nutritive value’ and ‘palatability’ did not

determine if a plant was immediately recalled and frequently named as a valuable forage

15

plant. By relating weighted ranks to criteria recorded with ecological methods, we were able

to identify and rank their relative importance. The attribution of a plant to a specific pasture

type was the most important non-explicit criterion for its valuation, followed by the plant’s

lifetime and its availability. Plants mainly growing on summer pastures had significantly

higher CSIanthro values than plants on all other pasture types. This was partly determined by

the higher average lifetime of plants growing on summer pastures, and by their higher

abundance and frequency (i.e. availability). However, even if the effects of lifetime and

availability on CSIanthro were statistically removed, the plant’s attribution to a certain pasture

type remained an important criterion for its local valuation. Plants growing on summer

pastures are much more valued than other plants.

Our data indicate that herdsmen used – besides the pasture type where a plant mainly occurs –

the lifetime and the availability of plants as indicators of their quality. The high valuation of

persistent species is in congruence with the ecological apparency hypothesis (EAH) and

corresponds to studies from elsewhere, showing that woody plant species have generally

higher use values than herbaceous plants (e.g. Ayantunde et al., 2009; Lucena et al., 2007). In

contrast to our results, though, earlier ethnobotanical studies found no positive correlation

between the relative importance and the availability of forage species (Lucena et al., 2007;

Lucena et al., 2012). Hence, the present study is to our knowledge the first to confirm the

validity of EAH within the use category of forage species. For Ait Toumert herdsmen,

visibility (expressed as lifetime) and availability (expressed as CSIeco values) were almost

equally important in determining a forage plant’s apparency.

However, herders had (besides the EAH criteria ‘visibility’ and ‘availability’) another implicit

criterion for the quality of a forage plant, i.e. its occurrence on upland pastures. We assume

that the need to cope with the high spatio-temporal variability of natural resources which is

typical for dryland systems (Thomas et al., 2007) may explain why apparent forage plants

from pasture types with a relatively constant environment were higher valued than those from

more arid environments. While persistent plants from upland pastures were hardly more

productive than those of lowland pastures (Linstädter and Baumann, 2013), they offered

reliable forage resources. We thus identify the reliability of dryland plants and pastures as a

key element for local valuation. Moroccan pastoral-nomads benefit from a cultural domain of

forage resources including reliability as a specific value in a highly variable environment.

Hence, reliability is - besides the two EAH criteria visibility and availability - another

important criterion to explain the use of plant resources in local communities. It also gives a

more functional explanation than EAH criteria why a woody forage plant should be more

valued than a herbaceous plant: It is not only more visible for livestock and/or for herdsmen,

but also offers more predictable forage resources.

As part of the interviews was carried out on summer pastures, our result that plants mainly

growing on summer pastures had significantly higher CSIanthro values could also be related to

informants’ bias (i.e. informants could have tended to name first species occurring on the

pasture type where they stayed during the time of interview). For logistical reasons, we were

not able to spread interviews over the whole annual cycle, which would have been necessary

to completely avoid possible bias. However, we assume any bias to be small, as ca. 50% of

interviews were carried out when households stayed on intermediate pastures. Another

methodological issue is the comparatively small sample size of free-lists (17 informants

representing more than 50 % of all Ait Toumert households) used for CSIanthro calculation. It

lies below the minimum number of 20-30 which is accepted as the norm to attain a robust

frequency parameter from free-list tasks (Sutrop, 2001). However, as mentioned in section

2.2, smaller samples may be sufficient if not only item frequency but also the mean position

of items is considered (as done in our study via CSI calculation). Sutrop (2001:272) explicitly

mentions this as an advantage of the CSI. Moreover, a methodological study on free-list tasks

16

– which evaluated the relative importance of sample size and the deletion of low-frequency

terms for robust frequency-based results – stressed the crucial importance of the latter

procedure only (Schrauf and Sanchez, 2010). Due to the fact that we omitted low-frequency

items prior to CSI calculation, and due the robustness of CSI with respect to a small sample

size, we assume that our conclusions based on CSIanthro are sound.

We triangulated our conclusions in April 2008 with an expert group interview with herdsmen

from different households. Herdsmen agreed that a plant with a higher lifetime would be a

more valuable forage resource, because it had a higher ability to survive drought times, and to

provide forage in bad and very bad years. They argued that plants growing on summer and

intermediate pastures were generally more valuable because they were a forage resource

which could be exploited annually.

We assume that plant valuation is amplified by the high cultural importance of reliable upland

pastures, which are under an exclusive tenure regime with strong institutions. In contrast,

winter pastures are an unreliable, low-quality resource type (following Scoones, 1999). They

are less contested and only protected by weak institutions. Hence, resource valuation may be a

self-enforcing process, with reliable, contested resources protected by strong institutions

becoming even more valuable in local perception. Incidentally, in the case of forage plants

EAH does not allow distinction whether a plant’s cultural importance is influenced by an

observation of livestock behaviour, and/or by herdsmen demonstrating the same behaviour as

herbivores (as assumed by Phillips and Gentry, 1993a, b). However, this conceptual

vagueness does not undermine the general applicability of EAH to forage plants. Instead of

use values, which were originally proposed within the context of EAH (Phillips and Gentry,

1993a, b), we have applied a combination of free-list rankings and weighted rank calculations

to obtain a quantitative measure of the relative importance of species. This approach has the

advantage that it can evaluate the relative importance within a single use category, and has for

this reason been recommended ‘if one wants to find the most culturally salient plants of a

particular sort’ (Quinlan, 2005). If one was more interested in explicit criteria for the

valuation of species, one could combine weighted rank calculations with explicit rankings

obtained in individual or group interviews (see e.g. Oba et al., 2008).

4.2. Linking the availability and valuation of forage resources to management decisions

Herdsmen’s reports on available resources were reasonably consistent for the past 3-4 years

but less so for years further in the past; for years more a decade away herdsmen tended to

recall only drought years. This appears to be common for pastoral and agro-pastoral societies

without written historical records or documentation about resources use (Meze-Hausken,

2004, Mogotsi et al., 2013). Precipitation records from the past six years corroborate the

accuracy of local reports for the six years, and suggest that the yardstick for an ‘average’ year

roughly corresponds to a ‘near-normal’ annual precipitation. Interestingly, the uncertainty

and risk experienced during winter months dominated the perception of past years.

Pastoral perception of forage resources on local pastures was an important component of

management decisions (Wesuls and Lang, 2010). In our case, Ait Toumert decision-making

can both be linked to the resource availability in a particular year and to the reliability of

forage plants on pasture types, i.e. to a non-explicit criterion for plant valuation. Pastures with

a high abundance of perennial plants were more reliable and thus more valuable. The reliable

upland pastures were regularly grazed during spring, summer and autumn. Herdsmen were

conscious about the unreliability of winter pastures where annual plant species are dominant.

Their decision to move there largely depended on the resource availability in a particular year.

An analysis of individual herd mobility patterns, based on mobility decision in four years

from the near past differing in resource availability, showed that considerable between-year

differences occurred in mobility decisions during winter time. Only in a year with good

17

rainfall, most of the households decided to move to winter pastures, because herdsmen knew

that they would then find an abundance of herbaceous biomass for their livestock. In other

types of years, only a certain proportion of pastoral-nomadic households moved to winter

pastures. Instead, households stayed on intermediate pastures where perennial species are

more abundant. Thus, mobility decisions depended on the availability of forage resources in

general (which is related to the amount of rainfall in a given year; Ruppert et al., 2012), and

specifically on the availability of forage resources provided from perennial species: The less

perennial species occurred on a certain pasture, the more the movement to this pasture

depended on the type of year.

The mobility patterns of the Ait Toumert allow an exploitation of the diversity of vegetation

types along a steep aridity gradient, securing the access to natural resources which were

highly variable in space and time. This opportunistic mobility, which implies ‘to be at the

right place at the right moment’ has been described as analogous to the movements of large

wild herbivores in their arid environments and would then be a mere coping or resource

exploitation strategy on a short-term level (Fabricius et al., 2007), i.e. a fall-back mechanism.

However, several aspects of Ait Toumert mobility patters may also be interpreted as a

resilient form of natural resource management. First, the regular seasonal movement between

pastures leaves enough forage on certain pasture types for the scarce times of the year or for

drought times (Linstädter et al., 2010), increasing forage resource reliability. Another

mechanism to promote resource reliability is through resting highly utilised pastures in years

with abundant rainfall (Müller et al., 2007a). In our case study, herds were frequently moved

to near and far winter pastures in years with good resource availability (Fig. 5), allowing

perennial vegetation on intermediate pastures to recover (Linstädter and Baumann, 2013). On

upland pastures, resource reliability is also maintained by means of mobility patterns

regulated by the agdal institution. The agdal protects perennial species on summer pastures

from grazing during the onset of the vegetation period (Genin and Simenel, 2011), when plant

individuals are highly sensitive to defoliation. This mechanism is clearly articulated by Ait

Toumert herders: ‘Forage plants need this time to flourish and to provide sufficient forage

during the summer months. If pastoral-nomads would move to the Awjgal [summer pasture]

in spring time, there would be less forage for livestock’. (Informant 14, October 6th

2007). The

agdal thus allows a recovery of palatable perennial plants, and minimizes resource

degradation via a deferred mobility. The crucial importance of protecting or improving

pasture reliability through resting in times when forage plants are particularly sensitive to

grazing has also been reported to build resilience in other dryland systems (Müller et al.,

2007b).

In arid and semi-arid environments with their high spatial and temporal variability, ecological

reliability is thus both sought after and protected via mobility patterns. Perceiving reliability,

and adjusting management decisions to it, are thus important mechanisms of resilience in

dryland pastoral systems (Fig. 6): They increase the system’s ability to cope with and

regenerate from external shocks such as droughts (Dougill et al., 2010).

Linking the local valuation of forage plants to mobility decisions has certain limitations as

individual mobility strategies are also influenced by socio-economic factors like individual

access to incorporated, monetary and socially institutionalized resources (Akasbi et al., 2012;

Breuer, 2007). Hence, the capacity of different households to search and protect ecological

reliability can vary (Kuhn et al., 2010).

4.3. New management strategies of the Ait Toumert: Implications for resilience

Today, Ait Toumert households are not as dependent on forage resources and their reliability

as they used to be. Feed supplementation became a common practice (Akasbi et al., 2012).

18

Transporting livestock to other grazing areas, and mass selling of animals are also new

management strategies. As typical for recent changes in pastoral societies of the High Atlas

region (Breuer, 2007) and elsewhere (Bollig, 2006; Greiner, 2011), new strategies such as

reciprocal grazing agreements with neighboring fractions, alternative monetary income from

wage labor, and socio-economic reciprocity have been introduced, and the social, economic,

or personal situation of a household influences its portfolio of old and new strategies, and its

selection of a herd management strategy in a specific situation. Alternative strategies also

seek or create reliability within the social-ecological system by buffering resource variability

(McAllister et al., 2006). However, the Ait Toumert increasingly rely not on ecological, but

on economic buffers: A resilience mechanism closely connected to the ecological system is

increasingly replaced by other mechanisms. Due to a weakening of environmental feedback

(Lorent et al., 2009; McAllister et al., 2006), the recent tendency of the Ait Toumert to

become decoupled from their natural resource base may accelerate resource degradation, and

decrease the resilience of the pastoral system. To evaluate the resilience of changing dryland

social-ecological systems, an integrated view of the social and natural sciences will be

necessary (Easdale and Aguiar, 2012).

Fig. 6. Interactions in a dryland pastoral system. Between the ecological subsystem (right-

hand side) and the social subsystem (left-hand side) of the social-ecological system, dynamic

interactions are the perception of natural resources, a decision-making-process leading to a

selection of a particular management strategy, and the impact of land management on

resources. Local perception of available natural resources is related to local ecological

knowledge on their spatio-temporal patterns and their quality.

5. Conclusion

Our study confirms the validity of EAH for the use category of forage species and for dryland

environments. EAH criteria ‘visibility’ and ‘availability’ explained the valuation of forage

plants even better than herdsmen’s explicit criteria. Our results also imply that EAH should be

modified for dryland environments with their high environmental variability. Apparent forage

plants from pastures within a relatively constant environment were higher valued than those

growing under more variable environmental conditions. We have introduced the concept of

reliability into EAH to explain the disproportionately high valuation of plants from less arid

19

environments. We could also show that reliability is a useful concept to link the perception of

natural resources in arid and semi-arid environments to mobility decisions. Finally, our results

underline that exploring local criteria for the valuation of natural resources can improve our

understanding of management decisions and resilience mechanisms in dryland systems.

Acknowledgements

We are deeply indebted to the Ait Toumert families for their hospitality and willingness to

share their knowledge. Our Moroccan partners, particularly Sabah Benchaaboune, Aziz

Rahou, and Ahmet Ramedane (CBTHA) are thanked for fruitful discussions on local

knowledge and rangeland management. Special thanks to Redouane Oumouch and Aziz

Labdi for translation and field assistance. We thank Jenny Eisold, Pierre Fritzsche and Silke

Tönsjost for technical and methodological support. Michael Bollig, Christoph Zielhofer, the

editor and two anonymous referees provided insightful comments on earlier versions of the

manuscript.

Disclosure statement

No actual or potential conflicts of interest have influenced the work of the authors. Authors

have closely collaborated in all stages of the research, and have approved the final version of

this manuscript.

Role of the funding source

This work was supported by the Federal German Ministry of Education and Research

(BMBF) under grant No. 01 LW 06001B, and by the Ministry of Innovation, Science,

Research and Technology (MIWFT) of the federal state of Northrhine-Westfalia under grant

No. 313-21200200. A. L. also acknowledges funding by the German Science Foundation

(DFG) through the collaborative research project FOR 1501. Funding sources have not

participated in, or conditioned any stage of the work here presented.

References

Akasbi, Z., Oldeland, J., Dengler, J., Finckh, M., 2012. Social and ecological constraints on

decision making by transhumant pastoralists: a case study from the Moroccan Atlas

Mountains. Journal of Mountain Science 9, 307-321.

Albuquerque, U.P., Lucena, R.F.P., 2005. Can apparency affect the use of plants by local

people in tropical forests? Interciencia 30, 506-510.

Ayantunde, A.A., Hiernaux, P., Briejer, M., Udo, H., Tabo, R., 2009. Uses of local plant

species by agropastoralists in south-western Niger. Ethnobotany Research & Applications 7,

53-66.

Baumont, R., Prache, S., Meuret, M., Morand-Fehr, P., 2000. How forage characteristics

influence behaviour and intake in small ruminants: a review. Livestock Production Science

64, 15-28.

Benabid, A., Fennane, M., 1994. Connaissances sur la végétation du Maroc: Phytogéographie,

phytosociologie et séries de végétation. Lazaroa 14, 21-97.

Bernard, H.R., 1994. Research Methods in Anthropology: Qualitative and Quantitative

Approaches. Sage Publications, Thousand Oaks, Canada.

Bollig, M., 2006. Risk Management in a Hazardous Environment - a Comparative Study of

two Pastoral Societies. Springer, New York.

20

Bollig, M., Schulte, A., 1999. Environmental change and pastoral perceptions: degradation

and indigenous knowledge in two African pastoral communities. Human Ecology 27, 493-

514.

Borgatti, S., 1996. Anthropac, 4.0 ed. Analytic Technologies, Lexington, USA.

Borgatti, S., 1999. Elicitation techniques for cultural domain analysis, in: Schensul, J.,

LeCompte, M., Nastasi, B.K., Borgatti, S. (Eds.), Enhanced Ethnographic Methods:

Audiovisual Techniques, Focused Group Interviews, and Elicitation Techniques. AltaMira

Press, Walnut Creek, USA, pp. 115-151.

Born, K., Fink, A.H., Paeth, H., 2008. Dry and wet periods in the northwestern Maghreb for

present day and future climate conditions. Meteorologische Zeitschrift 17, 533-551.

Breuer, I., 2007. Livelihood security and mobility in the High Atlas Mountains, in: Gertel, J.,

Breuer, I. (Eds.), Pastoral Morocco: Globalizing Scapes of Mobility and Insecurity. Reichert,

Wiesbaden, Germany, pp. 165-179.

Brewer, D.D., 2002. Supplementary interviewing techniques to maximize output in free

listing tasks. Field Methods 14, 108-118.

Butt, B., 2010. Seasonal space-time dynamics of cattle behavior and mobility among Maasai

pastoralists in semi-arid Kenya. Journal of Arid Environments 74, 403-413.

de Almeida, C.F.C.B., de Lima e, S., de Amorim, E.L.C., Maia, M.B., de Albuquerque, U.P.,

2005. Life strategy and chemical composition as predictors of the selection of medicinal

plants from the caatinga (Northeast Brazil). Journal of Arid Environments 62, 127-142.

Dougill, A.J., Fraser, E.D.G., Reed, M.S., 2010. Anticipating vulnerability to climate change

in dryland pastoral systems: using dynamic systems models for the Kalahari. 15(2). Ecology

and Society 15 (2), 17. [online] URL: http://www.ecologyandsociety.org/vol15/iss12/art17/.

Dwyer, M., Istomin, K., 2008. Theories of nomadic movement: A new theoretical approach

for understanding the movement decisions of Nenets and Komi reindeer herders. Human

Ecology 36, 521-533.

Easdale, M.H., Aguiar, M.R., 2012. Regional forage production assessment in arid and semi-

arid rangelands – A step towards social–ecological analysis. Journal of Arid Environments 83,

35-44.

Fabricius, C., Folke, C., Cundill, G., Schultz, L., 2007. Powerless spectators, coping actors,

and adaptive co-managers: a synthesis of the role of communities in ecosystem management.

Ecology and Society 12(1), 29. [online] URL:

http://www.ecologyandsociety.org/vol12/iss21/art29/.

Fazey, I., Proust, K., Johnson, B., Fazey, J.A., 2006. Eliciting the implicit knowledge and

perceptions of on-ground conservation managers of the Macquarie Marshes. Ecology and

Society 11, 25. [online] URL: http://www.ecologyandsociety.org/vol11/iss21/art25/.

Fernandez-Gimenez, M.E., 2000. The role of Mongolian nomadic pastoralists' ecological

knowledge in rangeland management. Ecological Applications 10, 1318-1326.

Finckh, M., Poete, P., 2008. Vegetation map of the Drâa basin. In: Schulz, O., Judex, M.

(Eds.), IMPETUS Atlas Morocco. Research Results 2000-2007. Department of Geography,

University of Bonn, Bonn, pp. 31–32.

Genin, D., Simenel, R., 2011. Endogenous Berber forest management and the functional

shaping of rural forests in Southern Morocco: implications for shared forest management

options. Human Ecology 39, 257-269.

Greiner, C., 2011. Migration, translocal networks and socio-economic stratification in

Namibia. Africa 81, 606-627.

21

Iav, 2002. Étude relative à l'inventaire participatif de la biodiversité dans le versant sud du

Haut Atlas, in: Institut Agronomique et Vétérinaire, H., II (Ed.). Conservation de la

Biodiversité par la Transhumance dans le Versant sud du Haut Atlas (CBTHA), Ouarzazate,

Morocco, pp. 1-78.

Johnson, J.B., Omland, K.S., 2004. Model selection in ecology and evolution. Trends in

Ecology & Evolution 19, 101-108.

Komwihangilo, D.M., Sendalo, D.S.C., Lekule, F.P., Mtenga, L.A., Temu, V.K., 2001.

Farmers' knowledge in the utilisation of indigenous browse species for feeding of goats in

semi arid central Tanzania. Livestock Research for Rural Development 13, 1-7.

Kuhn, A., Heidecke, C., Roth, A., Goldbach, H., Burkhardt, J., Linstädter, A., Kemmerling,

B., Gaiser, T., 2010. The importance of resource management for livelihood security under

climate change, in: Speth, P., Christoph, M., Diekkrüger, B. (Eds.), Impacts of Global Change

on the Hydrological Cycle in West and Northwest Africa. Springer, Berlin, pp. 565-590.

Lawrence, A., Phillips, O.L., Ismodes, A.R., Lopez, M., Rose, S., Wood, D., Farfan, A.J.,

2005. Local values for harvested forest plants in Madre de Dios, Peru: Towards a more

contextualised interpretation of quantitative ethnobotanical data. Biodiversity and

Conservation 14, 45-79.

Linstädter, A., Baumann, G., 2013. Abiotic and biotic recovery pathways of arid rangelands:

Lessons from the High Atlas Mountains, Morocco. CATENA 103, 3-15.

Linstädter, A., Baumann, G., Born, K., Diekkrüger, B., Fritzsche, P., Kirscht, H., Klose, A.,

2010. Land use and land cover in Southern Morocco: managing unpredictable resources and

extreme events, in: Speth, P., Christoph, M., Diekkrüger, B. (Eds.), Impacts of Global Change

on the Hydrological Cycle in West and Northwest Africa. Springer, Berlin, pp. 611-632.

Lorent, H., Sonnenschein, R., Tsiourlis, G.M., Hostert, P., Lambin, E., 2009. Livestock

subsidies and rangeland degradation in central Crete. Ecology and Society 14(2), 41. [online]

URL: http://www.ecologyandsociety.org/vol14/iss42/art41/.

Lucena, R.F.P., Araújo, E.L., Albuquerque, U.P., 2007. Does the local availability of woody

Caatinga plants (Northeastern Brazil) explain their use value? Economic Botany 61, 347-361.

Lucena, R.F.P., Medeiros, P.M., Araújo, E.L., Alves, A.G.C., Albuquerque, U.P., 2012. The

ecological apparency hypothesis and the importance of useful plants in rural communities

from Northeastern Brazil: An assessment based on use value. Journal of Environmental

Management 96, 106-115.

McAllister, R.R.J., Gordon, I.J., Janssen, M.A., Abel, N., 2006. Pastoralists' responses to

variation of rangeland resources in time and space. Ecological Applications 16, 572-583.

McKee, T.B., Doesken, N.J., Kleist, J., 1993. The relationship of drought frequency and

duration to time scales. 8th Conference on Applied Climatology, January 17-22, 1993,

Anaheim, California, pp. 179-184.

Meze-Hausken, E., 2004. Contrasting climate variability and meteorological drought with

perceived drought and climate change in northern Ethiopia. Climate Research 27, 19-31.

Middelton, N.J., Thomas, D.S.G., 1997. World Atlas of Desertification. United Nations

Environmental Programme (UNEP), London, UK.

Mogotsi, K., Nyangito, M.M., Nyariki, D.M., 2013. The role of drought among agro-pastoral

communities in a semi-arid environment: The case of Botswana. Journal of Arid

Environments 91, 38-44.

Müller, B., Frank, K., Wissel, C., 2007a. Relevance of rest periods in non-equilibrium

rangeland systems - a modelling analysis. Agricultural Systems 92, 295-317.

22

Müller, B., Linstädter, A., Frank, K., Bollig, M., Wissel, C., 2007b. Learning from local

knowledge: modeling the pastoral-nomadic range management of the Himba, Namibia.

Ecological Applications 17, 1857-1875.

Oba, G., Byakagaba, P., Angassa, A., 2008. Participatory monitoring of biodiversity in East

African grazing lands. Land Degradation & Development 19, 636-648.

Oldeland, J., Finckh, M., Born, K., 2008. A bioclimatic map for southern Morocco. In:

Schulz, O., Judex, M. (Eds.), IMPETUS Atlas Morocco. Research Results 2000–2007.

Department of Geography, University of Bonn, Bonn, pp. 21–22.

Phillips, O., Gentry, A., 1993a. The useful plants of Tambopata, Peru: I. Statistical hypotheses

tests with a new quantitative technique. Economic Botany 47, 15-32.

Phillips, O., Gentry, A., 1993b. The useful plants of Tambopata, Peru: II. Additional

hypothesis testing in quantitative ethnobotany. Economic Botany 47, 33-43.

Quinlan, M., 2005. Considerations for collecting freelists in the field: examples from

ethnobotany. Field Methods 17, 219-234.

Ruppert, J.C., Holm, A.M., Miehe, S., Muldavin, E., Snyman, H.A., Wesche, K., Linstädter,

A., 2012. Meta-analysis of rain-use efficiency confirms indicative value for degradation and

supports non-linear response along precipitation gradients in drylands. Journal of Vegetation

Science 23, 1035-1050.

Schrauf, R.W., Sanchez, J., 2010. Age effects and sample size in free listing. Field Methods

22, 70-87.

Scoones, I., 1999. Ecological dynamics and grazing-resource tenure: a case study from

Zimbabwe, in: Niamir-Fuller, M. (Ed.), Managing Mobility in African Rangelands. The

Legitimization of Transhumance. Intermediate Technology Publications Ltd, London, pp.

217-235.

Steele, M.Z., Shackleton, C.M., 2010. Using local experts as benchmarks for household local

ecological knowledge: Scoring in South African savannas. Journal of Environmental

Management 91, 1641-1646.

Sutrop, U., 2001. List task and a cognitive salience index. Field Methods 13, 263-276.

Szabó, A.T., 2004. Definitions, concepts and methods in the ethnobotany of food plants, in:

Macbeth, H., MacClancy, J. (Eds.), Researching Food Habits: Methods and Problems.

Berghahn, Oxford, UK, pp. 29-40.

Thomas, D.S.G., Twyman, C., 2004. Good or bad rangeland? Hybrid knowledge, science, and

local understandings of vegetation dynamics in the Kalahari. Land Degradation &

Development 15, 215-231.

Thomas, D.S.G., Twyman, C., Osbar, H., Hewitson, B., 2007. Adaptation to climate change

and variability: farmer responses to intra-seasonal precipitation trends in South Africa.

Climatic Change 83, 301-322.

Thompson, E.C., Juan, Z., 2006. Comparative cultural salience: Measures using free-list data.

Field Methods 18, 398-412.

Trabut, L., 2006. Noms Indigènes des Plantes d'Afrique du Nord. Ibis Press, Paris, France.

Wesuls, D., Lang, H., 2010. Perceptions and measurements: the assessment of pasture states

in a semi-arid area of Namibia. Human Ecology 38, 305-312.

Whitfield, S., Reed, M.S., 2012. Participatory environmental assessment in drylands:

Introducing a new approach. Journal of Arid Environments 77, 1-10.

Zemmrich, A., Manthey, M., Zerbe, S., Oyunchimeg, D., 2010. Driving environmental factors

and the role of grazing in grassland communities: A comparative study along an altitudinal

gradient in Western Mongolia. Journal of Arid Environments 74, 1271-1280.