Tree dieback affects climate change mitigation potential of a dry afromontane forest in northern Ethiopia

Forest Ecology and Management 344 (2015) 73–83

Contents lists available at ScienceDirect

Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco

Tree dieback affects climate change mitigation potential of a dry afromontane forest in northern Ethiopia Mulugeta Mokria a,b,⇑,1, Aster Gebrekirstos a, Ermias Aynekulu a, Achim Bräuning b a b

World Agroforestry Centre (ICRAF), United Nations Avenue, P.O. Box 30677-00100, Nairobi, Kenya Institute of Geography, Friedrich-Alexander-University Erlangen-Nuremberg, Wetterkreuz 15, 91058 Erlangen, Germany

a r t i c l e

i n f o

Article history: Received 3 December 2014 Received in revised form 6 February 2015 Accepted 9 February 2015 Available online 2 March 2015 Keywords: Allometric model Carbon sequestration Climate change Juniperus procera Olea europaea Dendrochronology

a b s t r a c t Extreme climatic events such as droughts are likely to result in huge and long-lasting effects on regional ecosystem health if large numbers of foundation tree species continue to die. Although deforestation is severe in the Ethiopian highlands, some remnants of dry afromontane forests still exist. However, the resilience of these forests under climate change scenarios is unknown. Therefore, we studied (1) the extent and spatial patterns of standing dead stems along an elevational gradient and (2) the effects of dieback on forest carbon sequestration potential and aboveground carbon stocks, in Juniperus procera and Olea europaea dominated dry afromontane forest in northern Ethiopia, using allometric models combined with tree ring analysis. Juniperus procera and Olea europaea constitute 67% of the total tree population. Tree dieback affected a quarter of the total population. This loss is critical because 92.2% of snags belong to J. procera and O. europaea, which are the foundation species of the study forest. The total estimated mean aboveground C-stock was 19.3 (±3.9) Mg C ha1. Of this estimate, snags contributed 34.5% of total C-stock. The estimated annual C-sequestration potential of the study forest was 0.33 (±0.03) Mg C ha1 year1, which is 27% less when compared to the pre-tree dieback C-sequestration potential. We found a decreasing trend in tree dieback with increasing elevation, which implies that the aboveground C-stocks and climate change mitigation potentials of the forest, was highly affected at lower elevations which is the drier part of the landscape. Tree ring analysis showed that trees reach medium-sized stem diameter (i.e., 20–25 cm) after no less than 100 years, indicating that the effect of forest dieback on C-sequestration potential and ecosystem function is long-lasting. Our results provide information on the magnitude of tree dieback and its long-lasting impact on forest carbon fluxes and forest ecosystem services. Evidently, the results substantiate the importance of protecting such forest to maintain the quality of the environment and to reduce efforts and cost for forest restoration after major loss. Finally, the information gained by this study provides baseline information for comparison of future carbon sequestration estimates. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Tropical forests play key roles in mitigating climate change (Saatchi et al., 2011). The reduction of tropical forest degradation and carbon emission from tropical forests have been proposed as one way to stabilize the atmospheric carbon cycle and to minimize impacts of future climate change (e.g., Corbera et al., 2010). However, climate change and anthropogenic stress factors are

⇑ Corresponding author at: World Agroforestry Centre (ICRAF), United Nations Avenue, P.O. Box 30677-00100, Nairobi, Kenya. Mobile: +254 708391541. E-mail addresses: [email protected], [email protected] (M. Mokria). 1 Present address: Institute of Geography, Friedrich-Alexander-University Erlangen-Nuremberg, Wetterkreuz 15, 91058 Erlangen, Germany. Tel.: +49 913185 29372, mobile: +49 15208390197. http://dx.doi.org/10.1016/j.foreco.2015.02.008 0378-1127/Ó 2015 Elsevier B.V. All rights reserved.

accelerating the rate of tropical forest degradation and increasing CO2 emission (Gibbs et al., 2007). In line with this, studies in Africa (Kherchouche et al., 2012), Asia (Nieuwstadt and Sheil, 2005), Australia (Fensham et al., 2009), Europe (Dobbertin et al., 2007), and America (Williams et al., 2010) have reported drought induced tree dieback, which diminishes ecosystem services that humans can obtain from forests (Gonzalez et al., 2012). Concerns about forest degradation and reduction in the associated ecosystem services are rising because the frequency of drought events, temperature increase, and heat waves are projected to increase in the future (IPCC, 2007). Particularly, areas that are already rather dry are the most likely to become even drier in the future (Dai, 2012), indicating that many dryland forests and woodland areas are more vulnerable to climate-induced forest dieback (Allen et al., 2010; Peng et al., 2011). In this regard, ongoing climate change might accelerate tree mortality especially in countries like Ethiopia,

74

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

where dryland areas cover 72% of the total land area (Lemenih and Kassa, 2011). The ecosystem services that can be generated from the dry afromontane forests of Ethiopia, are threatened due to forest degradation induced mainly by anthropogenic pressures, including extensive forest resources utilization and land use changes (Bongers et al., 2006; Tesfaye et al., 2014; Zeleke and Hurni, 2001). In the study forest, J. procera and O. europaea are dominant and so that they are regarded as foundation tree species (e.g., Ellison et al., 2005). Earlier studies conducted in the northern dry afromontane forests of Ethiopia have generated substantial information on population structure (Wassie et al., 2010), biodiversity (Aerts et al., 2006; Aynekulu et al., 2012), natural regeneration (Aynekulu et al., 2009), and soil seed bank (Lemenih and Teketay, 2006). However, these studies did not investigate the aboveground biomass and carbon dynamics, the extent of tree dieback and the associated loss of aboveground carbon and annual carbon sequestration potential of the remnant dry afromontane forest. As a result, the importance of the remnant dry afromontane forests of Ethiopia for climate change adaptation and mitigation and their resilience against increasing drought is hardly known. However, such information is critical to design forest conservation strategies that help reducing the impacts of climate change (Chidumayo et al., 2011). Thus, the objectives of this study were to (i) investigate the current extent of forest degradation due to tree dieback, (ii) quantify the effects of tree dieback on aboveground carbon stock and carbon sequestration potential of the study forest, and (iii) determine growth dynamics of foundation tree species. Based on former studies (e.g., Ellison et al., 2005, 2010) we hypothesized that declining of foundation tree species (e.g., J. procera and O. europaea) from the system may have long-lasting effects on forest fundamental ecosystem processes and services. 2. Methods and materials 2.1. Study area and forest This study was carried out in the Desa’a dry afromontane forest, located (13°360 to 13°560 N and 39°480 to 39°510 E) in the semiarid agro-ecological zone of Tigray region, northern Ethiopia (Fig. 1). The study area lies in the transition zone between the Acacia-Commiphora woodland and shrubland in the Afar lowlands (1400 m.a.s.l) and the dry evergreen afromontane forest and grassland complex in the Tigray highlands (2800 m.a.s.l). Topographically, the study area is mountainous and nearly 45% of the area has slope inclination greater than 30% (Aynekulu, 2011). The geological structure is diverse, playing a major role for soil variability. A large part of Desa’a forest is characterized by shallow soils and frequent rock outcrops of Enticho sandstone and Crystalline Basement (Asrat, 2002). The dominant soil types are Leptosols, Cambisols, Vertisols, Regosols, and Arenosols (Aynekulu, 2011). Desa’a forest is a dry afromontane forest dominated by Juniperus procera and Olea europaea subsp. cuspidatae. Some of the other co-occurring tree species (hereafter, other species) in the study forest are Solanum schimperianum Hochst. ex A. Rich, Dodonaea viscosa Jacq, Maytenus senegalensis (Lam.) Exell, Carissa edulis Vahl, Rhus sp. nov. A, and Clutia lanceolata Forssk, in decreasing order of abundance (Aynekulu et al., 2009). 2.2. Climate characteristics of the study area The regional climate shows a distinct seasonality in precipitation with a unimodal rainfall pattern. The peak rainy season lasts from July to August and the dry season from October to June. Based

on data obtained from the National Meteorological Agency of Ethiopia for the years 2006–2013, the mean annual rainfall at the nearest climate station (i.e., Atsibe station, 13°530 N; 39°440 E; 2711 m.a.s.l.) was 602.1 (±62.1) mm. The mean minimum and maximum monthly mean temperatures were 9.2 (±0.6) °C and 19.9 (±0.42) °C, respectively (Fig. 2). 2.3. Field measurements, sample collection and wood density measurement We established five transects perpendicular to the main slope, considering topographic variation, tree size and species composition. In total, 57 plots of 50 m  50 m size were set up at 100 m elevational intervals. In each plot, we measured biometric parameters such as stem diameter (D) at breast height (1.3 m above the ground) and tree height (H) for individual living trees and snags (i.e., standing dead trees) using diameter tape and clinometer, respectively. We documented the geographic coordinates of each plot and derived the elevation data from a topographic map (1:50,000). To minimize errors, we did not measure the height of strongly bent trees or snags that have lost the total crown area. The proportion of trees without height measurement was 20.5%, 21% and 28.3% for J. procera, O. europaea and other species, respectively. To determine the tree age and annual radial increment rate, we collected 20 stem discs from different living trees and snags of J. procera. The samples were transported to the dendrochronological laboratory of the World Agroforestry Center-(ICRAF), Nairobi, Kenya, and air-dried and sanded. To determine wood density, we used the stem discs harvested for ring analysis. We extracted wood samples of about five cm width from each disc that include the sapwood and heartwood parts of the discs. For wood density measurement, the samples were rehydrated for one day (MartínezCabrera et al., 2009) and the wood density was analyzed as described by Williamson and Wiemann (2010): a beaker with enough size to hold the sample was filled with water and placed on an electronic balance of 0.01 g precision. Then, the wood sample was carefully immersed in the water. The displacement weight by the sample was recorded and converted to sample fresh volume using the formula: displacement weight (g)/density of water at a standard temperature and pressure. Afterward, the samples were placed in the oven for 72 h at 105 °C to obtain dry mass. Wood density (q) of J. procera was calculated from dry weight to fresh volume ratio. For O. europaea and all other co-occurring tree species, values of q were derived from the global wood density database (Zanne et al., 2009). From this database, a species-specific wood density, 0.913 g/cm3 was obtained for O. europaea. For other species, an average wood density value of q = 0.58 g/cm3 calculated from 279 tropical Africa tree species was used. 2.4. Site-specific tree height-diameter allometry Tree height (H) is one of the most important predictor in quantifying aboveground forest biomass and carbon fluxes (Chave et al., 2014). Although, the inclusion of H as predictor can considerably improve the accuracy of forest aboveground biomass estimations (Chave et al., 2005; Feldpausch et al., 2012), measuring tree height is difficult, especially in a mixed-species tropical forest (Feldpausch et al., 2011). To overcome this problem, tree height is usually derived from regional or continental-scale H:D models. However, the H–D relationship may considerably vary between species, vegetation structures, geographical locations and due to climatic factors (Feldpausch et al., 2011; Osman et al., 2013). Hence, it is crucial to develop site-specific H:D relationships for accurate estimation of H from observed D, for improving the accuracy of local biomass estimates (Ketterings et al., 2001; Osman et al., 2013).

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

75

Fig. 1. Location of the study site.

24

250

19.9 (± 0.42) C 0

200

22

150 20 100 18

50

Temperature ( oC)

Precipitation (mm)

602.1(± 62.1)mm

function (Ketterings et al., 2001; Muller-Landau et al., 2006), for living trees of J. procera and O. europaea, and a third equation by pooling all other species together. These models were used to derive H for few individuals without H measurement from their observed D (Zhang et al., 2014).

H ¼ a  Db

ð1Þ

where H is height (m), D is diameter (cm), a is the scaling coefficient, and b is the scaling exponent. 2.5. Estimation of C-stock in above ground biomass

16

Dec

Oct

Nov

Sep

Jul

Aug

Jun

May

Apr

Mar

Jan

Feb

0

Months Fig. 2. Mean monthly rainfall (mm, black bars) and mean maximum temperature (°C, black line) at Atsibe meteorological station for the period 2006–2013. Numbers provided indicate mean annual precipitation and mean maximum temperature.

There is no single and universally accepted scaling relationship between tree growth and size (Muller-Landau et al., 2006). Usually, non-linear functions, including the power-law function are used to construct predictive relationships between tree H and D (Zhang et al., 2014). A power-law relationship between tree height and diameter is biologically feasible, particularly for trees with diameters <30 cm) (Muller-Landau et al., 2006; Osman et al., 2013). Therefore, by considering our data range (see Table 1), we constructed site-specific H:D allometric equations using a power-law

The destructive above ground tree biomass (AGB) estimation method is the most recommended and accurate (Clark and Kellner, 2012; Phuong et al., 2012), although it is laborious and inapplicable on a wider geographic scales (Pilli et al., 2006). As an alternative, biomass allometric equations are used to estimate forest aboveground biomass and carbon stock (Chave et al., 2014). In recent decades, several allometric models were developed for a variety of tree species and forest types (Henry et al., 2013). Unfortunately, these models are not evenly distributed across regions. For example, species-specific allometric equations are available for only 1% of tree species in Sub-Saharan Africa (SSA) (Henry et al., 2011). Given this limited geographical distribution of biomass estimation models, the accuracy of AGB estimates still falls behind the required quality (Clark and Kellner, 2012; Henry et al., 2013). Moreover, Sileshi (2014) has discussed common statistical mistakes encountered during developing allometric models, which introduce potential sources of uncertainty in biomass estimation. Even

76

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

Table 1 Summary of plot inventory data including living trees and snags of a dry afromontane forest in northern Ethiopia. DBH, H, and SE refer to diameter at breast height, tree height, and standard error, respectively. Species

Status

No. of trees

Mean [SE] DBH (cm)

Mean [± SE] H (m)

Range of DBH (cm)

Range of H (m)

5–15

15–30

30–50

>50

J. procera

alive snag

1069 607

16.5 [0.96] 17.2 [1.17]

6.1 [0.29] 5.9 [0.29]

5–88 5–90

2–17.5 2–20

60.6 48.4

33.8 44.2

4.8 5.6

0.8 1.8

O. europaea

alive snag

1313 802

18.5 [0.83] 19.6 [1.23]

5.5 [0.16] 4.7 [0.16]

5–90 5–114

2–17 2–13.5

39.2 48.3

52.7 46.4

7.0 3.7

1.1 1.6

Other species

alive snag

1747 120

11.4 [0.43] 10.5 [0.73]

4.2 [0.09] 3.6 [0.20]

5–85 5–27

2–17 2–8.4

78.1 74.2

20.8 22.5

1.0 0.0

0.1 3.3

though, the choice of the most accurate allometric model is not straightforward (Henry et al., 2013), and it is necessary to assess the nature of existing allometric models prior to application (Melson et al., 2011; Sileshi, 2014). In order to reduce uncertainty due to model selection, authors have suggested some basic parameters of models that have to be checked while choosing biomass estimation models (for details see: (Chave et al., 2005, 2014; Henry et al., 2011; Ketterings et al., 2001; Melson et al., 2011; Sileshi, 2014). Considering these suggestions, we selected two biomass estimation models to convert inventory data into AGB and Cstock. The first allometric model is the improved pan-tropical mixedspecies biomass estimation model by Chave et al. (2014):

AGB1 ¼ a  ðq  D2  HÞ

b

ð2Þ

 with coefficients a = 0.0673 and b = 0.976 and parameters AGBest, in kilogram, q in gram per cubic centimeter, D in centimeter, and H in meter. The second model is the flexible tropical mixed-species biomass estimation model by Ketterings et al. (2001):

AGB2 ¼ r  q  D2þb

et al., 2011). To understand the magnitude of imprecision in measured H, authors have followed a re-measurement approach and reported different values of imprecision. For instance, Chave et al. (2004) and Hunter et al. (2013) reported imprecision of 10% and 12% in measured H, respectively. Hence, we used 10% uncertainty in the measured tree H for error propagation (Chave et al., 2004; Schöngart et al., 2011). To account for the uncertainties of q in the estimated biomass, we used the standard deviation of q for J. procera (Schöngart et al., 2011). In case of O. europaea, we used an uncertainty of 10% in q, based on Chave et al. (2004) who reported 10% imprecision in measured wood density. For all other species without species-specific information of q, we apply the standard deviation of the mean qfrom 279 tropical Africa tree species (Chave et al., 2004; Schöngart et al., 2011). The uncertainty in the calculation of r is assumed to be 10%. For Eq. (2), we applied the following algorithm to propagate measurement errors in rD, rH, rq and rr, to estimated AGB:

h

r2 AGBð1Þ ¼ a  b  qb1  rp  ðH  D2 Þ

i b 2

h i b 2 þ a  b  Hb1  rH  ðq  D2 Þ h i2 b1 þ a  b  ðD2 Þ  rD  ðq  HÞb

ð3Þ

 with coefficients D in centimeter, q in gram per cubic centimeter, AGBest, in kilogram, r is a constant parameter over a range of sites calculated as r = a/q(wood specific gravity), where a = 0.066, is the constant parameter, b is a scaling exponent derived from species-specific height-diameter allometry using (Eq. (1)). In each plot, we estimated AGB for each individual tree using Eqs. (2) and (3) and extrapolated estimations to a per-hectare (Mg ha1). The carbon content in the above ground biomass was estimated by multiplying the values of AGB by the default IPCC carbon fraction value of 0.47 (IPCC, 2006). However, the conversion of forest inventory data into AGB and C-stock still has some degree of uncertainty arising from measurements errors (e.g., D, H, and q), sample-plot size selection and the choice of biomass estimation models (Chave et al., 2004; Larjavaara and Muller-Landau, 2013; Sileshi, 2014). To account for the relative contribution of measurement errors in the total uncertainty of AGB and C-stock estimates, we performed propagation of measurement errors for individual trees (Chave et al., 2004; Schöngart et al., 2011). Then the total uncertainties were extrapolated to a per-hectare basis (Mg C ha1). Although stem diameter (D) measurement is not problematic, the landscape of the study site and some of multi-stemmed, knotted and branched trees might affect the measurements. Thus, authors have reported different magnitudes of imprecision in measured D, e.g., 2% (Fahey and Knapp, 2007), 5% (Chave et al., 2004) and 5.3% (Shimizu et al., 2014), respectively. Considering these reports and the topography of the study site, we used 5% uncertainty in the measured D. It is well known that tree height (H) measurement, particularly, in a mixed-species forest is difficult (Feldpausch

Proportion of trees (%) under diff. diameter class (cm)

ð4Þ

For Eq. (3) the following algorithm was applied to calculate measurement errors propagation:

h

i2

h

r2 AGBð2Þ ¼ rr  q  Db þ r  rq  Db

i2

h i2 þ r  q  rD  b  ðDÞb1

ð5Þ

From Eqs. (4) and (5), we estimated propagation of error for each model separately. We also calculated the proportion of each error to the total uncertainty in the estimated AGB for J. procera, O. europaea, and for all other species separately. Then, the total uncertainty for each plot by accumulating all variances was calculated (Eq. (6)).

rAGBð1Þ ¼

X

r2AGBð1Þ

0:5

and

rAGBð2Þ ¼

X

r2AGBð2Þ

0:5

ð6Þ

Then, the mean AGB and new variance were calculated form Eqs. (2) and (3) to consider errors due to model selection using the following formula (Schöngart et al., 2011).

h

i

r2AGBðmeanÞ ¼ r2AGBð1Þ þ r2AGBð2Þ  0:25

ð7Þ

2.6. Tree ring analysis to estimate C-sequestration rate in the AGB A successful implementation of a carbon–offsetting project needs high-resolution and larger scale quantitative data on carbon

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

stock and annual carbon accumulation rate in the forest. However, it is often difficult to acquire such information at larger spatial scales, because of huge resource requirement. The shortage of data on forest aboveground carbon stocks have been partially solved through application of biomass estimation models (Chave et al., 2014), although estimating the annual carbon sequestration rate is still challenging. Recently, dendrochronology is regarded as a robust, rapid and cost-effective tool for forest carbon accounting with temporal resolution (Gebrekirstos et al., 2014; Schöngart et al., 2011). The tree ring analysis was conducted for J. procera. For this analysis, we followed standard dendrochronological procedures (Cook and Kairiukstis, 1990). The dendrochronological potential of J. procera has been demonstrated by other studies conducted in Ethiopia (Sass-Klaassen et al., 2008; Wils et al., 2010). To improve the visibility of growth-ring boundaries, the cross-sectional surfaces of the stem disks were air-dried and progressively sanded using finer grades of sanding paper of grit-sizes from 60 to 1200. The wood anatomy and features that characterize growth ring boundaries were studied from transversal micro-thin sections with a thickness of 20–30 lm. The sanded stem discs were inspected under a stereomicroscope to detect and mark concentric growth ring boundaries. Then, ring widths were measured to the nearest 0.01 mm using a digital measuring device (LINTAB, RINNTECH, Inc.) supported by the software TSAP (Time Series Analyses and Presentation, Rinntech, Germany). To understand the long-term growth trajectory of J. procera, the mean annual diameter increments (MAI) were calculated for each individual tree from annual ring width using: MAI = cumulative growth (CGW(t))/year (t) (Schöngart et al., 2007). This biometric parameter has been used to account for forest biomass/C-stocks, and carbon sequestration rate (Krause, 2010; van Laar and Akça, 2007; West, 2009). A predictive relationship between tree age obtained from tree ring analysis and stem diameter (D) was conducted using a non-linear, sigmoidal regression function:

D¼

a 1 þ expððage  cÞ=bÞ

ð8Þ

In order to examine the relationship between tree age and C-stock in AGB, we conducted a non-linear regression analysis between tree age obtained from tree ring analysis and its estimated C-stock by applying Eqs. (2) and (3):

C-stock ¼

a 1 þ expððage  cÞ=bÞ

ð9Þ

From cumulative C-stock we modeled annual carbon sequestration (kg year1 tree1) using the following formula: C-sequestration = C-stock(t+1)  C-stock(t). The annual C-sequestration potential of J. procera in the study forest was quantified by multiplying the values of the annual C-sequestration rate obtained from the model (Eq. 9) and living tree stem density of J. procera per-plot. Then, the estimates were extrapolated to a per-hectare basis (Mg C ha1 yr1). To account for the total C-sequestration potential of the study forest, we first quantified the annual C-sequestration potential for O. europaea and other species based on the annual C-sequestration rate estimated for J. procera. Afterward, the estimates conducted for J. procera, O. europaea and other species, were summed together and reported as a per-hectare basis (Mg C ha1 yr1). To evaluate the impacts of forest dieback on C-sequestration potential of foundation species and in general, the pre-tree-dieback was calculated from the annual C-sequestration rate, and multiplied by the total stem density (i.e. including living and snag trees). Finally, we deducted the value estimated for living trees (i.e., posttree-dieback).

77

2.7. Statistical analysis We tested the mean differences among the plots in tree diameter, height, and C-stock in the AGB using the Kruskal–Wallis test (nonparametric one-way ANOVA). The mean diameter differences between living trees and snags were tested using an independent t-test. To inspect the relationships between wood density, age, and annual diameter increment of J. procera, Pearson correlation tests were carried out. We also conducted a non-linear regression analysis to examine the relationship between tree age and its diameter and C-stock in the above ground biomass. A Pearson correlation analyses were conducted to test the relationship between elevation and stem density. Finally, we tested for linear relationships between stem density and predicted relative uncertainties in the estimated C-stock in the AGB. 3. Results 3.1. Stem density, size and extent of tree dieback The two dominant tree species J. procera and O. europaea comprised about 67% of the total number of the recorded individual trees (Table 1). However, the frequency of these two foundation tree species differs considerably between the studied plots, ranging from 0–90% and 0–68% for J. procera and O. europaea, respectively. Stem density varied between 68 and 1112, with the overall mean of 397 (±31) trees per hectare. In the study forest, there are few larger trees (Table 1). The number of larger trees with DBH > 50 cm of J. procera, O. europaea and all other species constitute 1.2%, 1.3%, and 0.11% of the total population, respectively. However, the frequency of stem size distribution revealed that the diameter class ranging from 5 to 30 cm contained a large proportion of trees (Table 1). We found a significant difference between plots in average stem diameter (p < 0.001) and tree height (p < 0.001). The diameter and height of J. procera and O. europaea were larger compared to other species (Table 1). We also detected significant H–D relationships for J. procera (n = 821, r2 = 0.5, p < 0.001), O. europaea (n = 987, r2 = 0.5, p < 0.001), and all other species (n = 1259, r2 = 0.42, p < 0.001). The extent of tree dieback of both species decreased with increasing elevation suggesting that tree dieback is more sever in the lower (drier) part of the landscape. The proportion of snags in each plot varied between 0.7% and 82%. From the total recorded J. procera and O. europaea, about 37% were snags. This is as large as 92.2% of the total recorded snags in the studied forest. The average stem diameter of snags of J. procera and O. europaea is slightly higher than the diameter of living trees, although not significantly different (p > 0.05) (Table 1). 3.2. Growth dynamics and wood characteristics The microscopic analysis of the wood anatomy of J. procera showed distinct growth ring boundaries. The wood formed during one annual growth period can be distinguished by thin-walled tracheids in early wood and thick-walled tracheids in the late wood (Fig. 3). The flattened and rectangular tracheids at the end of the growth ring demarcate growth ring boundaries. J. procera showed different growth rates over time (Fig. 3). Tree ages determined from tree ring counting ranged between 111 and 262 years. The annual radial increment varied between 0.25 and 2.10 mm year1, with the overall mean of 0.91 (±0.02) mm year1. On average, J. procera requires at least 100 years to attain a diameter of 20– 25 cm (Fig. 4). The relationship between tree age and D is statically significant (r2 = 0.6, n = 20, P < 0.001). The mean wood density (q) of J. procera was 0.60 ± 0.07 g cm3. Wood density and tree age

78

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

Fig. 3. A micro-thin section photograph showing growth rates of J. procera varying over time. Solid arrows indicate growth ring boundaries and the white arrow indicates growth direction.

The proportions of aboveground C-stock estimated from snags were different between tree species (Table 2) and increased with diameter class (Fig. 5a–c). In the study forest, trees with stem diameter greater than 50 cm comprised 25%, 22%, and 7.4% of total aboveground C-stocks in O. europaea, J. procera and other species, respectively. Furthermore, snags from J. procera and O. europaea constitute 99% of the total aboveground C-stock estimated from snags in the study forest. Aboveground C-stock was strongly influenced by position along the altitudinal gradient. C-stocks were greater at higher altitudes (Fig. 6a and b), following a positive linear relationship in case of J. procera (r = 0.48, p < 0.05, n = 31) and O. europaea (r = 0.37, p < 0.05, n = 53). However, the relationship between altitude and C-stock in other species was negative (r = 0.45, p < 0.01, n = 57). We found a large proportion of C-stock fixed in snags at lower altitudes (Fig. 6a–c). 3.4. Dry afromontane forest C-sequestration potential

0.24

50

0.18

30 20

0.12

MAI (cm)

CGW (cm)

40

CGW MAI

10

0.06

0 0

50

100

150

200

250

Forest dieback considerably affected C-sequestration potential of the study forest, though the impact is significantly higher on foundation tree species (Table 3). The mean annual C-sequestration rate of individual J. procera trees was 1.12 (±0.05) kg C tree1 year1. The tree ring analysis of J. procera revealed a continuous carbon sequestration for about 200 years (Fig. 7). Moreover, total carbon sequestered estimated by Eqs. (2) and (3) was significantly correlated with tree age (r2 = 0.70, p < 0.001, n = 20). The mean Csequestration potential of J. procera was 0.14 (±0.03) Mg C ha1 year1 (Table 3). The overall mean annual C-sequestration potential of the studied dry afromontane forest was 0.33 (±0.03) Mg C ha1 year1. However, this estimate was lower when compared to the pre-tree dieback C-accumulation potential (Table 3). 3.5. Propagated uncertainties to estimated aboveground C-stock

Years (Age)

were positively correlated (r = 0.7, p < 0.01, n = 20), and a negative correlation was found between wood density and mean annual diameter increment (r = 0.64, p < 0.01, n = 20).

The average uncertainties in the estimated aboveground Cstock of J. procera, O. europaea and all other species were higher when using Eq. (3) (Table 4). We found a significant (r = 0.52, p < 0.01, n = 57) negative relationship between relative uncertainty in the estimates and stem density. The total uncertainties due to measurement errors in the estimates were 14.4% in Eq. (2) and 20.2% in Eq. (3).

3.3. C-stock in the above ground biomass and spatial distributions

4. Discussion

Trees considerably varied in their C-stocks in AGB. Individual trees from J, procera and O. europaea store more carbon than trees of all other species. However, individual snags of J. procera store more carbon than living trees (Table 2). Sample plots were significantly (P < 0.001) different in aboveground C-stocks. The mean estimated aboveground C-stock from Eqs. (2) and (3) was 19.3 (±3.9) Mg C ha1 (Table 2). Furthermore, J. procera and O. europaea alone stored more than 90% of carbon in the study forest (Table 2).

4.1. Stem density, tree size and growth dynamics

Fig. 4. Cumulative diameter growth (CGW) and mean annual diameter increment (MAI) of J. procera from a dry afromontane forest in northern Ethiopia.

Tree size (i.e. D and H) variation across the studied plots implies that microsite conditions, such as soil physical and chemical characteristics play a significant role for tree growth. Species composition and forest structure might also have influenced tree growth at a sub-plot level. The study forest is an old-growth forest, but more than 90% of foundation tree species have DBH between 5 and

Table 2 Estimated aboveground C-stock for a dry afromontane forest in northern Ethiopia, derived from forest inventory data using allometric models. Total uncertainties of estimates are presented in brackets. Species

J. procera O. europaea Other species All species

Mean aboveground C-stock (kg C tree1)

Total aboveground C-stock (Mg C ha1)

Living trees

Snags

Eq. (2)

Eq. (3)

Mean

Proportion of C-stock estimated from snags (%)

41.0 80.7 22.1 43.6

57.8 72.9 12.2 62.0

8.3 (±1.5) 11.0 (±2.2) 2.0 (±0.5) 17.2 (±3.5)

10.6 (±1.8) 13.9 (±2.6) 2.4 (±0.7) 21.4 (±4.3)

9.4 (±1.6) 12.4 (±2.4) 2.2 (±0.6) 19.3 (±3.9)

44.5 35.6 3.7 34.5

(±7.7) (±17.6) (±6.6) (±9.9)

(±10.9) (±15.9) (±3.6) (±14.1)

79

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

(a)

28

J. procera

10 8

(b) O. europaea

24

Living trees Snags

Carbon stock (Mg)

6 4 2

Living trees Snags

20 16 12 8 4

Carbon stock (Mg)

8

(c)

>50

45-50

40-45

35-40

30-35

25-30

20-25

5-10

Diameter class (cm)

15-20

0

>50

45-50

40-45

30-35

35-40

25-30

20-25

15-20

5-10

10-15

0

10-15

Carbon stock (Mg)

12

Diameter class (cm)

Other species Living trees Snags

6

4

2

>50

45-50

40-45

35-40

30-35

25-30

20-25

15-20

10-15

5-10

0

Diameter class (cm) Fig. 5. Total aboveground carbon stock under different diameter classes of living trees and snags in a dry afromontane forest in northern Ethiopia.

Carbon stock (Mg)

14

(a)

20 J. procera Living trees Snags

12 10

Carbon stock (Mg)

16

8 6 4

(b)

O. europaea Living trees Snags

15

10

5

2 0 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 2700 2800

1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 2700 2800

0

Elevation (m.a.s.l)

Elevation (m.a.s.l)

Carbon stock (Mg)

6

(c)

Other species Living trees Snags

4

2

1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 2700 2800

0

Elevation (m.a.s.l) Fig. 6. Total aboveground carbon stock in living trees and snags along an elevation gradient in a dry afromontane forest in northern Ethiopia.

80

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

Table 3 Estimated (mean ± SE) annual C-sequestration potential for a dry afromontane forest in northern Ethiopia. CS refers to carbon sequestration. Mean annual C-sequestration potential (Mg C ha1 year1)

Species

J. procera O. europaea Other species All species

Pre-tree-dieback

Post-tree dieback

Lost CS-potential (%)

0.22 0.18 0.15 0.45

0.14 0.11 0.14 0.33

36.4 39.0 6.7 27.0

(±0.03) (±0.02) (±0.01) (±0.03)

(±0.03) (±0.02) (±0.01) (±0.03)

30 cm, indicating that they have been growing slowly. This may be partially explained by a tradeoff between resource availability-uptake rate and survival strategy of the species (Gebrekirstos et al., 2006, 2011; Sterck et al., 2006). Besides, the old age, but relatively small size trees underline the general stress level for tree growth at the study site. The average radial increment rate was less than reported by SassKlaassen et al. (2008), suggesting that J. procera responds strongly to local environmental and climatic conditions. Several dendrochrological studies conducted in central Ethiopia (Krepkowski et al., 2011; Sass-Klaassen et al., 2008), in the central rift valley of Ethiopia (Gebrekirstos et al., 2008), in west Africa (Gebrekirstos et al., 2012) and in other tropical regions (e.g., Worbes, 2002) found that rainfall seasonality is the main growth-driving factor. In Ethiopia, species growing in areas with unimodal rainfall pattern show annual growth ring formation in response to rainfall seasonality. Examples are J. procera (e.g., Wils et al., 2010), and Boswellia papyrifera in northern Ethiopia (Tolera et al., 2013), and several species in the Acacia-dominated woodland in the central rift valley (Gebrekirstos et al., 2008). Since the study site experiences a unimodal rainfall regime with only two months with rainfall above 100 mm (Fig. 2), it is legitimate to postulate that the anatomical growth ring boundaries found in J. procera at our study site in general represent annual growth rings. More importantly, due to the low annual precipitation and high evapotranspiration at the study site, it is reasonable to assume that climate change over time has been affecting radial growth rates of foundation tree species. Future dendroclimatological research is required to substantiate this conclusion.

Table 4 Propagation of measurement errors for converting forest inventory data (D, H, q, and r) into estimated aboveground C-stock using allometric models.

4.2. The magnitude of dieback of foundation tree species The existence of snags in all studied plots indicate that common external stress factors, such as increase in temperature, reduction in rainfall amount and increase in frequency of extreme drought events, may have been inducing tree mortality along the whole

0.24

O. europaea Total error rD rH

2.0 0.18

1.5 1.0

0.12 0.5 0.0

C-sequestration MAI

0

50

100

150

200

250

J. procera Total error rD rH

rq rr

MAI (cm)

C-sequestration (Kg C year -1)

2.5

altitudinal gradient. In this remnant dry afromontane forest, more than a quarter of the tree population had died, suggesting that tree dieback is a critical problem. The significance of this problem becomes evident from the fact that 92.2% of snags are from foundation tree species, that eventually affects their structural and functional characteristics which is crucial to form habitat for a large number of related tree species and define ecosystem processes, such as, energy and nutrient fluxes (Ellison et al., 2005). Even though both J. procera and O. europaea are drought-tolerant trees (e.g., Bussmann, 2006), a multi-year cumulative stress may aggravate poor tree growth and have increased susceptibility to extreme drought events (Suarez and Ghermandi, 2004). Moreover, the projected extreme drought events may induce further growth rate reduction which is likely to increase mortality rates (Carus, 2010; Suarez and Ghermandi, 2004). The mortality rate of O. europaea is higher than that of J. procera. This may be related to the fact that J. procera is predominantly found in the more mesic high-elevation parts of the landscape, whereas O. europaea dominates the lower elevations which are more exposed to the influence of heat waves originating in the adjacent Denakil depression with an active basaltic shield volcano (Erta Ale). Hence, the increasing number of snags with decreasing altitude implies that position within the altitudinal gradient contributes to the mortality rates of trees. According to local key informants (elders), besides increasing temperature and reduction of rainfall, an increase in the frequency of heat waves from the Denakil depression is one of the main causes for tree dieback in the study area. Forest dieback due to extreme drought has been reported from different geographical locations in the tropics (e.g., Allen et al., 2010; Kherchouche et al., 2012). A study in Ethiopia (Aynekulu et al., 2011) found that the forest-woodland ecotone had moved about 500 m slope upwards and pioneer tree species are taking over at lower elevations. The absence of foundation tree species at lower elevation suggests that drought-tolerant species are taking over the region where seedlings from foundation species can no longer survive. A study by Krepkowski et al. (2013) using intra-annual analysis of carbon isotopes, showed differences in carbon allocation between early-successional pioneer species and

rq rr Other species Total error rD rH

rq rr 0.06

Years (Age) Fig. 7. Estimated annual carbon sequestration rate and mean annual diameter increment (MAI) of J. procera in relation to tree age.

All species Total error rD rH

rq rr

Eq. (2)

Eq. (3)

12.1% 1.2% 57.1% 41.7% –

18.0% 41.4% – 24.7% 33.9%

13.8% 0.9% 49.6% 49.5% –

18.9% 38.9% – 30.5% 30.5%

16.8% 1.2% 41.2% 57.6% –

26.0% 38.3% – 36.0% 25.7%

14.4% 1.0% 50.0% 49.0% –

20.2% 39.4% – 29.9% 30.7%

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

81

late-successional conifers and describe implications on plant–soil– atmosphere carbon balances of different forest successional stages, which have to be considered for carbon management of tropical forest landscapes. Moreover, the loss of this forest may expose the surrounding agricultural lands to heat waves from the Denakel depression and this ultimately may affect soil moisture status through increased evapotranspiration (Bonan, 2008) and may end up with reducing agricultural production. To substantiate the results of findings, future studies should focus on assessing the temporal patterns of forest dieback by collecting stem disks from snags along the whole altitudinal gradient either using radio-carbon dating or in combination with dendrochronology. 4.3. Estimated C-stock in AGB and spatial distributions Our results underestimate the total aboveground C stock of the study forest, as we did not consider trees with a diameter of less than 5 cm. The different estimates of C-stock in AGB from Eqs. (2) and (3) suggested that allometric differences and predictors included in the models might play a considerable role in the estimation of aboveground C-stock from forest inventory data. For example, reduction values of AGB were found in Eastern and Western Africa after including height (H) as predictor (Feldpausch et al., 2012). Estimated AGB in this study was similar to biomass estimated for Western African community managed forests (Skutsch and Ba, 2010) and Western African agroforestry parklands (Luedeling and Neufeldt, 2012). Moreover, our result was in the range of C-stocks estimated for tropical dry forests (Gibbs et al., 2007). Differences in tree size, stem density, and the proportion of foundation tree species in different plots may have resulted in spatial variations of estimated aboveground C-stocks. We estimated a much higher C-stock from snags of J. procera and O. europaea, suggesting that this dry afromontane forest is under serious degradation. Beside, larger size snags are currently rotting and release carbon to the atmosphere, eventually changing the forest’s carbon balance from a C-sink status to a C-source. 4.4. Impacts of tree dieback on C-sequestration potential of dry afromontane forest Despite of our slight underestimation of the actual C-sequestration potential, the annual C-sequestration potential of the study forest is very low compared to the annual C-sequestration potential of other tropical forests. This is due to the slow growth rates of foundation tree species and low stem density accompanied with higher tree dieback. The amount of trimmed C-sequestration potential of the study forest mainly due to tree dieback was 27% (Table 3). This loss is enormous and the effect is long-lasting, because a considerable number of medium-sized trees, which are active in carbon sequestrations, are dead and due to slow growth rate, small trees reach the medium-sized level after no less than 100 years (Fig. 4). Even though foundation tree species (i.e. J. procera and O. europaea) grow slowly and accumulate little biomass annually, they may amass more carbon than fast growing tree species during their longer life span. More importantly, they can also store carbon for a long period due to their slow carbon turnover rate (Krepkowski et al., 2013; Sterck et al., 2006). 4.5. Propagation of uncertainties Measurement errors in D, H, and q, could potentially increase the uncertainties of biomass estimates from forest inventory data, but also vary between tree species and model used (Table 4). The application of the same value of wood density for all other trees species may have resulted in higher uncertainty in the estimates. This further supports Chave et al. (2005), who have recommended

Fig. 8. Schematic representation of forest management options based on the extent of tree dieback at a micro-site level and elevation gradient. Note: The dieback zones are classified based on foundation species (J. procera and O. europaea). FS indicate foundation species. At the lower-left corner, the rectangular and circular area depicts implementation of targeted-exclosures with different size and shape on the most affected micro-sites within the forest. The superscript numbers on species names represent the rank based on abundance and the species to be centered, while implementing a given forest management options along elevational gradient.

wood density values to be included as a predictor in the calculation of aboveground biomass (in the order of importance): speciesspecific, family-specific, average wood density calculated for trees grown in a similar region. Among the predictors included in the model, imprecision of tree height measurements resulted in substantial uncertainties in aboveground C-stock estimates. Hence, we suggest that great caution needs to be taken when using tree height data generated from regional as well as global heightdiameter allometries to convert inventory data into biomass estimates. Furthermore, our finding is congruent with other studies (Chave et al., 2004; Hunter et al., 2013; Schöngart et al., 2011) who reported that measurement errors in (rD, rH, and rq) magnify uncertainties in the estimated aboveground C-stock. Therefore, considering measurement errors in the predictors (i.e. D, H, and q) may substantiate the reliability of estimated aboveground C-stock from forest inventory data. 4.6. Implications for forest management and restoration The extent of dieback considerably varied (i.e. 0.7–82%) along elevation gradient, indicating that micro-site conditions play a significant role for tree growth and survival in the study forest. In addition, the results indicate that the restoration or forest management activities should emerge from the understanding of the extent of dieback at micro-site scale, within the forest. This might help to prioritize on the most affected parts of the forest and take targeted measures to restore the area. The slow-growth rates observed in this study would imply the presence of resource limitations that are important for tree growth. This in turn urges the need to characterize specific site conditions, while planning and implementing forest management and restoration interventions. We also suggest considering successful restoration practice in the region, while planning and implementing forest restoration interventions. In particular, we suggest an ecosystem-based approach to restore the degraded forests. Specific management options are summarized in Fig. 8. 5. Conclusions Forest dieback had resulted in reduction of ecosystem functions and regulating ecosystem services (e.g., C-sequestration potential)

82

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83

that can be obtained from forest resources in the study forest. Such negative impacts might be long-lasting, as more than 90% of dead trees belong to foundation tree species and that they are slowgrowing trees. This in turn indicates that rehabilitation of this forest would require huge efforts. We suggest that rehabilitation efforts for this forest should first target at protecting the remaining trees, as they are the main driver for natural regeneration and speeding the ecosystem process rates. Secondly, measures should integrate enrichment planting, particularly, at the lower altitude. Finally, efforts should encourage/assist the community to generate income from this forest non-destructively. The information generated in this study will enlighten the impacts of forest degradation on forest ecosystem services and carbon fluxes. Furthermore, it is useful as baseline for future aboveground C-stock inventories, and helps to derive research-based forest development, conservation and monitoring strategies to combat the impacts of climate change and to build up climate variability resilient landscapes. Acknowledgements We are grateful to Edith Anyango for her assistant during the lab work. We are particularly grateful to Dr. Emiru Birhane for his facilitation to the research work, and to Mekele University for logistical support. We also thank the local community in the study area for their support during fieldwork. We are grateful to FTA-CRP 6.4 for funding the study. References Aerts, R., Van-Overtveld, K., Haile, M., Hermy, M., Deckers, J., Muys, B., 2006. Species composition and diversity of small Afromontane forest fragments in northern Ethiopia. Plant Ecol. 187, 127–142. http://dx.doi.org/10.1007/s11258-0069137-0. Allen, C.D., Macalady, A.K., Chenchouni, H., Bachelet, D., McDowell, N., Vennetier, M., Kitzberger, T., Rigling, A., Breshears, D.D., Hogg, E.H. (Ted), Gonzalez, P., Fensham, R., Zhang, Z., Castro, J., Demidova, N., Lim, J.-H., Allard, G., Running, S.W., Semerci, A., Cobb, N., 2010. A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests. For. Ecol. Manage. 259, 660-684. doi: http://dx.doi.org/10.1016/j.foreco.2009.09.001. Asrat, A., 2002. The rock-hewn churches of Tigrai, Northern Ethiopia: a geological perspective. Geoarchaeology 17, 649–663. http://dx.doi.org/10.1002/gea.10035. Aynekulu, E., 2011. Forest diversity in fragmented landscapes of northern Ethiopia and implications for conservation. A PhD thesis, Rheinschen Friedrich-Wilhems Universitat, Bonn, Germany. Aynekulu, E., Denich, M., Tsegaye, D., 2009. Regeneration Response of Juniperus procera and Olea europaea subsp cuspidata to Exclosure in a Dry Afromontane Forest in Northern Ethiopia. Mt. Res. Dev. 29, 143–152. http://dx.doi.org/ 10.1659/mrd.1076. Aynekulu, E., Denich, M., Tsegaye, D., Aerts, R., Neuwirth, B., Boehmer, H.J., 2011. Dieback affects forest structure in a dry Afromontane forest in northern Ethiopia. J. Arid Environ. 75, 499–503. http://dx.doi.org/10.1016/ j.jaridenv.2010.12.013. Aynekulu, E., Aerts, R., Moonen, P., Denich, M., Gebrehiwot, K., Vågen, T.-G., Mekuria, W., Boehmer, H.J., 2012. Altitudinal variation and conservation priorities of vegetation along the Great Rift Valley escarpment, northern Ethiopia. Biodivers. Conserv. 21, 2691–2707. http://dx.doi.org/10.1007/s10531-012-0328-9. Bonan, G.B., 2008. Forests and climate change: forcings, feedbacks, and the climate benefits of forests. Science 320, 1444–1449 (80-). Bongers, F., Wassie, A., Sterck, F.J., Bekele, T., Teketay, D., 2006. Ecological restoration and church forests in northern Ethiopia. J. Dryl. 1, 35–44. Bussmann, R.W., 2006. Vegetation zonation and nomenclature of African Mountains – an overview. Lyonia 11, 41–66. Carus, S., 2010. Pre-growth mortality of Abies cilicica trees and mortality models performance. J. Environ. Biol. 31, 363–368. Chave, J., Condit, R., Aguilar, S., Hernandez, A., Lao, S., Perez, R., 2004. Error propagation and scaling for tropical forest biomass estimates. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359, 409–420. http://dx.doi.org/10.1098/rstb.2003.1425. Chave, J., Andalo, C., Brown, S., Cairns, M.A., Chambers, J.Q., Eamus, D., Fölster, H., Fromard, F., Higuchi, N., Kira, T., Lescure, J.-P., Nelson, B.W., Ogawa, H., Puig, H., Riéra, B., Yamakura, T., 2005. Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145, 87–99. http://dx.doi.org/ 10.1007/s00442-005-0100-. Chave, J., Réjou-Méchain, M., Búrquez, A., Chidumayo, E., Colgan, M.S., Delitti, W.B., Duque, A., Eid, T., Fearnside, P.M., Goodman, R.C., Henry, M., Martínez-Yrízar, A., Mugasha, W.A., Muller-Landau, H.C., Mencuccini, M., Nelson, B.W., Ngomanda, A., Nogueira, E.M., Ortiz-Malavassi, E., Pélissier, R., Ploton, P., Ryan, C.M.,

Saldarriaga, Vieilledent, G., 2014. Improved allometric models to estimate the aboveground biomass of tropical trees. Glob. Chang. Biol, doi: http://dx.doi.org/ 10.1111/gcb.12629. Chidumayo, E., Okali, D., Kowero, G., Larwanou, M., 2011. Climate Change and African Forest and Wildlife Resources in African Forestry. African Forest Forum, Nairobi, Kenya. Clark, D.B., Kellner, J.R., 2012. Tropical forest biomass estimation and the fallacy of misplaced concreteness. J. Veg. Sci. 23, 1191–1196. http://dx.doi.org/10.1111/ j.1654-1103.2012.01471.x. Cook, E.R., Kairiukstis, L.A. (eds), 1990. Methods of dendrochronology: applications in the environmental sciences. Dordrecht, The Netherlands: Kluwer Academic Publishers and International Institute for Applied Systems Analysis, pp. 394. Corbera, E., Estrada, M., Brown, K., 2010. Reducing greenhouse gas emissions from deforestation and forest degradation in developing countries: revisiting the assumptions. Clim. Change 100, 355–388. http://dx.doi.org/10.1007/s10584009-9773-1. Dai, A., 2012. Increasing drought under global warming in observations and models. Nat. Clim. Chang. 3, 52–58. http://dx.doi.org/10.1038/nclimate1633. Dobbertin, M., Wermelinger, B., Bigler, C., Bürgi, M., Carron, M., Forster, B., Gimmi, U., Rigling, A., 2007. Linking increasing drought stress to Scots pine mortality and bark beetle infestations. Sci. World J. 7 (Suppl 1), 231–239. http:// dx.doi.org/10.1100/tsw.2007.58. Ellison, A.M., Bank, M.S., Clinton, B.D., Colburn, E.A., Elliott, K., Ford, C.R., Foster, D.R., Kloeppel, B.D., Knoepp, J.D., Lovett, G.M., Mohan, J., Orwig, D.A., Rodenhouse, N.L., Sobczak, W.V., Stinson, K.A., Stone, J.K., Swan, C.M., Thompson, J., Von Holle, B., Webster, J.R., 2005. Loss of foundation species: consequences for the structure and dynamics of forested ecosystems. In a nutshell. Front. Ecol. Env. 3, 479–486. Ellison, A.M., Barker-Plotkin, A.A., Foster, D.R., Orwig, D.A., 2010. Experimentally testing the role of foundation species in forests: the Harvard Forest Hemlock Removal Experiment. Methods Ecol. Evol. 1, 168–179. http://dx.doi.org/ 10.1111/j.2041-210X.2010.00025.x. Fahey, T.J., Knapp, A.K. (Eds.), 2007. Principles and Standards for Measuring Primary Production. Oxfored University Press. Inc. 198, Madison Avenue, New York, USA. Feldpausch, T.R., Banin, L., Phillips, O.L., Baker, T.R., Lewis, S.L., Quesada, C.A., AffumBaffoe, K., Arets, E.J.M.M., Berry, N.J., Bird, M., Brondizio, E.S., de Camargo, P., Chave, J., Djagbletey, G., Domingues, T.F., Drescher, M., Fearnsid, P.M., Franca, M.B., Fyllas, N.M., Lopez-Gonzalez, G., Hladik, A., Higuchi, N., Hunter, M.O., Iida, Y., Salim, K.A., Kassim, A.R., Keller, M., Kemp, J., King, D.A., Lovett, J.C., Marimon, B.S., Marimon-Junior, B.H., Lenza, E., Marshall, A.R., Metcalfe, D.J., Mitchard, _ S., E.T.A., Moran, E.F., Nelson, B.W., Nilus, R., Nogueira, E.M., Palace, M., Patino, Peh, K.S.H., Raventos, M.T., Reitsma, J. M., Saiz, G., Schrodt, F., Sonke, B., Taedoumg, H.E., Tan, S., White, L., Wöll, H., Lloyd, J., 2011. Height-diameter allometry of tropical forest trees. Biogeosciences 8, 1081–1106. doi: 10.5194/ bg-8-1081-2011. Feldpausch, T.R., Lloyd, J., Lewis, S.L., Brienen, R.J.W., Gloor, E., Monteagudo Mendoza, A., Lopez-Gonzalez, G., Banin, L., Abu Salim, K., Affum-Baffoe, K., Alexiades, M., Almeida, S., Amaral, I., Andrade, A., Aragão, L.E.O.C., Araujo Murakami, A., Arets, E.J.M.M., Arroyo, L., Baker, T.R., Bánki, O.S., Berry, N.J., Cardozo, N., Chave, J., Comiskey, J.A., Dávila, E.A., de Oliveira, A., DiFiore, A., Djagbletey, G., Domingues, T.F., Erwin, T.L., Fearnside, P.M., França, M.B., Freitas, M.A., Higuchi, N., Iida, Y., Jiménez, E., Kassim, A.R., Killeen, T.J., Laurance, W.F., Lovett, J.C., Malhi, Y., Marimon, B.S., Marimon-Junior, B.H., Lenza, E., Marshall, A.R., Mendoza, C., Metcalfe, D.J., Mitchard, E.T.A., Nelson, B.W., Nilus, R., Nogueira, E.M., Parada, A., Peh, K.S.-H., Pena Cruz, A., Peñuela, M.C., Pitman, N.C.A., Prieto, A., Quesada, C.A., Ramírez, F., Ramírez-Angulo, H., Reitsma, J.M., Rudas, A., Saiz, G., Salomão, R.P., Schwarz, M., Silva, N., Silva-Espejo, J.E., Silveira, M., Sonké, B., Stropp, J., Taedoumg, H.E., Tan, S., ter Steege, H., Terborgh, J., Torello-Raventos, M., van der Heijden, G.M.F., Vásquez, R., Vilanova, E., Vos, V., White, L., Wilcock, S., Woell, H., Phillips, O.L. 2012. Tree height integrated into pan-tropical forest biomass estimates. Biogeosciences 9, 2567–2622. doi: http://dx.doi.org/10.5194/bgd-9-2567-2012. Fensham, R.J., Fairfax, R.J., Ward, D.P., 2009. Drought-induced tree death in savanna. Glob. Chang. Biol. 15, 380–387. http://dx.doi.org/10.1111/j.13652486.2008.01718.x. Gebrekirstos, A., Teketay, D., Fetene, M., Mitlo, R., 2006. Adaptation of five cooccurring tree and shrub species to water stress and its implication in restoration of degraded lands. For. Ecol. Manage. 229, 259–267. http:// dx.doi.org/10.1016/j.foreco.2006.04.029. Gebrekirstos, A., Mitlöhner, R., Teketay, D., Worbes, M., 2008. Climate–growth relationships of the dominant tree species from semi-arid savanna woodland in Ethiopia. Trees 22, 631–641. http://dx.doi.org/10.1007/s00468-008-0221-z. Gebrekirstos, A., van Noordwijk, M., Neufeldt, H., Mitlohiner, R., 2011. Relationships of stable carbon isotopes, plant water potential and growth an approach to asses water use efficiency and growth strategies of dry land agroforestry species. Tree 25, 95–102. http://dx.doi.org/10.1007/s00468-010-0467-0. Gebrekirstos, A., Mitlöhner, R., Van Noordwijk, M., Bräuning, A., 2012. Tracing responses to climate variability from stable isotopes in tree rings of Anogeissus leiocarpus and Sclerocarya birrea from the Sahel zone, Burkina Faso. TRACE 10, 6–12. Gebrekirstos, A., Bra, A., Sass-klassen, U., Mbow, C., 2014. Opportunities and applications of dendrochronology in Africa. Curr. Opin. Environ. Sustain. 48–53. http://dx.doi.org/10.1016/j.cosust.2013.10.011. Gibbs, H.K., Brown, S., Niles, J.O., Foley, J.A., 2007. Monitoring and estimating tropical forest carbon stocks: making REDD a reality. Environ. Res. Lett. 2, 045023. http://dx.doi.org/10.1088/1748-9326/2/4/045023.

M. Mokria et al. / Forest Ecology and Management 344 (2015) 73–83 Gonzalez, P., Tucker, C.J., Sy, H., 2012. Tree density and species decline in the African Sahel attributable to climate. J. Arid Environ. 78, 55–64. http://dx.doi.org/ 10.1016/j.jaridenv.2011.11.001. Henry, M., Picard, N., Trotta, C., Manlay, R.J., Valentini, R., Bernoux, M., Saint-andré, L., 2011. Estimating tree biomass of sub-Saharan African forests: a review of available allometric equations. Siliva Fenn. 45 (3B), 477–569. Henry, M., Bombelli, A, Trotta, C., Alessandrini, A, Birigazzi, L., Sola, G., Vieilledent, G., Santenoise, P., Longuetaud, F., Valentini, R., Picard, N., Saint-André, L., 2013. GlobAllomeTree: international platform for tree allometric equations to support volume, biomass and carbon assessment. iForest – Biogeosci. For. 6, 326–330. doi: 10.3832/ifor0901-006. Hunter, M.O., Keller, M., Victoria, D., Morton, D.C., 2013. Tree height and tropical forest biomass estimation. Biogeosciences 10, 8385–8399. http://dx.doi.org/ 10.5194/bg-10-8385-2013. IPCC, 2006. Agriculture, forestry and other land use. Chapter 4. In: Eggleston, S., Beundia, L., Miaw, K., Ngara, T., Tanabe, K., editors. Forest land. 2006 IPCC guidelines for national greenhouse inventories, vol. 4. Hayama, Japan: Intergovernmental Panel on Climate Change, I. IPCC, 2007. Summary for Policymakers. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Ave). Ketterings, Q.M., Coe, R., van Noordwijk, M., Ambagau’, Y., Palm, C.A., 2001. Reducing uncertainty in the use of allometric biomass equations for predicting above-ground tree biomass in mixed secondary forests. For. Ecol. Manage. 146, 199–209. http://dx.doi.org/10.1016/S0378-1127(00)00460-6. Kherchouche, D., Kalla, M., Gutiérrez, E.M., Attalah, S., Bouzghaia, M., 2012. Impact of Droughts on Cedrus atlantica Forests Dieback in the Aurès (Algeria). J. Life Sci. 6, 1262–1269. Krause, M., 2010. Project on Environmental Protection and Management of Natural Resources in Dak Nong Province (EPMNR) Vietnam: Carbon Budget and Project Model V. 1.1. Humburg, Germany. Krepkowski, J., Bräuning, A., Gebrekirstos, A., Strobl, S., 2011. Cambial growth dynamics and climatic control of different tree life forms in tropical mountain forest in Ethiopia. Trees 25, 59–70. http://dx.doi.org/10.1007/s00468-0100460-7. Krepkowski, J., Gebrekirstos, A., Shibistova, O., Br, A., 2013. Stable carbon isotope labeling reveals different carry-over effects between functional types of tropical trees in an Ethiopian mountain forest. New Phytol. 199, 431–440. http:// dx.doi.org/10.1111/nph.12266. Larjavaara, M., Muller-Landau, H.C., 2013. Measuring tree height: a quantitative comparison of two common field methods in a moist tropical forest. Methods Ecol. Evol. 4, 793–801. http://dx.doi.org/10.1111/2041-210X.12071. Lemenih, M., Kassa, H., 2011. Opportunities and Challenges for Sustainable Production and Marketing of Gums and Resins in Ethiopia. CIFOR, Bogor, Indonesia. Lemenih, M., Teketay, D., 2006. Changes in soil seed bank composition and density following deforestation and subsequent cultivation of a tropical dry Afromontane forest in Ethiopia. Trop. Ecol. 47, 1–12. Luedeling, E., Neufeldt, H., 2012. Carbon sequestration potential of parkland agroforestry in the Sahel. Clim. Change 115, 443–461. http://dx.doi.org/ 10.1007/s10584-012-0438-0. Martínez-Cabrera, H.I., Jones, C.S., Espino, S., Schenk, H.J., 2009. Wood anatomy and wood density in shrubs: responses to varying aridity along transcontinental transects. Am. J. Bot. 96, 1388–1398. http://dx.doi.org/10.3732/ajb.0800237. Melson, S.L., Harmon, M.E., Fried, J.S., Domingo, J.B., 2011. Estimates of live-tree carbon stores in the Pacific Northwest are sensitive to model selection. Carbon Bal. Manag. 6, 2. http://dx.doi.org/10.1186/1750-0680-6-2. Muller-Landau, H.C., Condit, R.S., Chave, J., Thomas, S.C., Bunyavejchewin, S.A., Bohlman, S., Davies, S., Foster, R., Gunatilleke, S., Gunatilleke, N., Harms, K.E., Hart, T., Hubbell, S.P., Itoh, A., Kassim, A.R., LaFrankie, J.V., Lee, H.S., Losos, E., Makana, J.-R., Ohkubo, T., Sukumar, R., Sun, I.-F., Nur Supardi, M.N., Tan, S., Thompson, J., Valencia, R., Muñoz, G.V., Wills, C., Yamakura, T., Chuyong, G., Dattaraja, H.S., Esufali, S., Hall, P., Hernandez, C., Kenfack, D., Kiratiprayoon, S., Suresh, H.S., Thomas, D., Vallejo, M.I., Ashton, P., 2006. Testing metabolic ecology theory for allometric scaling of tree size, growth and mortality in tropical forests. Ecol. Lett. 9, 575–588, 10.1111/j.1461-0248.2006.00904.x. Nieuwstadt, M.G.L.V.A.N., Sheil, D., 2005. Drought, fire and tree survival in a Borneo rain forest, East Kalimantan, Indonesia. J. Ecol. 191–201. http://dx.doi.org/ 10.1111/j.1365-2745.2005.00954.x. Osman, E.L.M.H., Idris, E.L.Z.A., Ibrahim, E.L.M.M., 2013. Modelling height-diameter relationships of selected economically important natural forests species. J. For. Prod. Ind. 2, 34–42. Peng, C., Ma, Z., Lei, X., Zhu, Q., Chen, H., Wang, W., Liu, S., Li, W., Fang, X., Zhou, X., 2011. A drought-induced pervasive increase in tree mortality across Canada’s

83

boreal forests. Nat. Clim. Chang. 1, 467–471. http://dx.doi.org/10.1038/ nclimate1293. Phuong, V.T., Inoguchi, A., Birigazzi, L., Henry, M., Sola, G., 2012. Introduction and Background of the study, Vietnam. In: Inoguchi, A., Henry, M., Birigazzi, L., Sola, G. (Eds.), Tree allometric equation development for estimation of forest aboveground biomass in Viet Nam (Part A), UN-REDD Programme, Hanoi, Viet Nam. pp. 1–20. Pilli, R., Anfodillo, T., Carrer, M., 2006. Towards a functional and simplified allometry for estimating forest biomass. For. Ecol. Manage. 237, 583–593. http:// dx.doi.org/10.1016/j.foreco.2006.10.004. Saatchi, S.S., Harris, N.L., Brown, S., Lefsky, M., Mitchard, E.T.A., Salas, W., 2011. Benchmark map of forest carbon stocks in tropical regions across three continents. PNAS 108, 9899–9904. http://dx.doi.org/10.1073/pnas.1019576108. Sass-Klaassen, U., Couralet, C., Sahle, Y., Sterck, F.J., 2008. Juniper from Ethiopia contains a large-scale precipitation signal. Int. J. Plant Sci. 169, 1057–1065. http://dx.doi.org/10.1086/590473. Schöngart, J., Wittmanna, F., Worbesb, M., Piedadec, M.T., Junk, H.-J.K.W.J., 2007. Management criteria for Ficus insipida Willd. (Moraceae) in Amazonian whitewater floodplain forests defined by tree-ring analysis. Ann. For. Sci. 64, 657– 664. Schöngart, J., Arieira, J., Felfili Fortes, C., Cezarine de Arruda, E., Nunes da Cunha, C., 2011. Age-related and stand-wise estimates of carbon stocks and sequestration in the aboveground coarse wood biomass of wetland forests in the northern Pantanal, Brazil. Biogeosciences 8, 3407–3421. http://dx.doi.org/10.5194/bg-83407-2011. Shimizu, A., Yamada, S., Arita, Y., 2014. Diameter measurements of the upper parts of trees using an ultra-telephoto digital photography system. Open J. For., 316– 326 Sileshi, G.W., 2014. A critical review of forest biomass estimation models, common mistakes and corrective measures. For. Ecol. Manage. 329, 237–254. http:// dx.doi.org/10.1016/j.foreco.2014.06.026. Skutsch, M.M., Ba, L., 2010. Crediting carbon in dry forests: the potential for community forest management in West Africa. For. Policy Econ. 12, 264–270. http://dx.doi.org/10.1016/j.forpol.2009.12.003. Sterck, F.J., Poorter, L., Schieving, F., 2006. Leaf traits determine the growth-survival trade-off across rain forest tree species. Chicago J. 167, 758–765. Suarez, M.L., Ghermandi, L., 2004. Factors predisposing episodic drought-induced tree mortality in Nothofagus – site, climatic sensitivity and growth trends. J. Ecol., 954–966 Tesfaye, S., Guyassa, E., Joseph Raj, A., Birhane, E., Wondim, G.T., 2014. Land use and land cover change, and woody vegetation diversity in human driven landscape of Gilgel Tekeze Catchment, Northern Ethiopia. Int. J. For. Res. 2014, 1–10. http://dx.doi.org/10.1155/2014/614249. Tolera, M., Sass-Klaassen, U., Eshete, A., Bongers, F., Sterck, F.J., 2013. Frankincense tree recruitment failed over the past half century. For. Ecol. Manage. 304, 65–72. http://dx.doi.org/10.1016/j.foreco.2013.04.036. van Laar, A., Akça, A., 2007. Forest Mensuration, Managing Forest Ecosystems. Springer Netherlands, Dordrecht. doi: 10.1007/978-1-4020-5991-9. Wassie, A., Sterck, F.J., Bongers, F., 2010. Species and structural diversity of church forests in a fragmented Ethiopian Highland landscape. J. Veg. Sci. 21, 938–948. http://dx.doi.org/10.1111/j.1654-1103.2010.01202.x. West, P.W., 2009. Tree and Forest Measurement. Springer, Berlin Heidelberg. http:// dx.doi.org/10.1007/978-3-540-95966-3. Williams, A.P., Allen, C.D., Millar, C.I., Swetnam, T.W., Michaelsen, J., Still, C.J., Leavitt, S.W., 2010. Forest responses to increasing aridity and warmth in the southwestern United States. PNAS 107, 21289–21294. http://dx.doi.org/ 10.1073/pnas.0914211107. Williamson, G.B., Wiemann, M.C., 2010. Measuring wood specific gravity . . . correctly. Am. J. Bot. 97, 519–524. http://dx.doi.org/10.3732/ajb.0900243. Wils, T.H.G., Sass-Klaassen, U.G.W., Eshetu, Z., Bräuning, A., Gebrekirstos, A., Couralet, C., Robertson, I., Touchan, R., Koprowski, M., Conway, D., Briffa, K.R., Beeckman, H., 2010. Dendrochronology in the dry tropics: the Ethiopian case. Tree 25, 345–354. http://dx.doi.org/10.1007/s00468-010-0521-y. Worbes, M., 2002. One hundred years of tree-ring research in the tropics ± a brief history and an outlook to future challenges. Dendrochronologia 2, 217–231. Zanne, A.E., Lopez-Gonzalez, G., Coomes, D.A., Ilic, J., Jansen, S., Lewis, S.L., Miller, R.B., Swenson, N.G., Wiemann, M.C., and Chave, J., 2009. Global wood density database. Dryad. Identifier: . Zeleke, G., Hurni, H., 2001. Implications of land use and land cover dynamics for mountain resource degradation in the Northwestern Ethiopian highlands. Mt. Res. Dev. 21, 184–191. Zhang, X., Duan, A., Zhang, J., Xiang, C., 2014. Estimating tree height-diameter models with the Bayesian method. ScientificWorldJournal 2014, 683691. http:// dx.doi.org/10.1155/2014/683691.