Jordan Soil Carbon Inventory - Farm, Table & Sky

1 downloads 0 Views 591KB Size Report
Predicted soil organic carbon stocks and changes in Jordan between ... b Faculty of Agriculture, Jordan University of Science and Technology, Irbid 22110, ...

Agriculture, Ecosystems and Environment 122 (2007) 35–45 www.elsevier.com/locate/agee

Predicted soil organic carbon stocks and changes in Jordan between 2000 and 2030 made using the GEFSOC Modelling System R. Al-Adamat a,*, Z. Rawajfih b, M. Easter c, K. Paustian c, K. Coleman d, E. Milne e, P. Falloon f, D.S. Powlson d, N.H. Batjes g a

Institute of Earth and Environment Sciences, Al al-Bayt University, P.O. Box 130334, Mafraq, Jordan b Faculty of Agriculture, Jordan University of Science and Technology, Irbid 22110, Jordan c The Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, CO 80523-1499, USA d The Agriculture and Environment Division, Rothamsted Research, Harpenden, Herts AL5 2JQ, UK e The Department of Soil Science, The University of Reading, P.O. Box 233, Reading RG6 6DW, UK f The Met. Office, Hadley Centre for Climate Prediction and Research, Fitzroy Road, Exeter EX1 3PB, UK g ISRIC - World Soil Information, P.O. Box 353, 6700 AJ Wageningen, The Netherlands Available online 15 February 2007

Abstract Estimates of soil organic carbon (SOC) stocks and changes under different land use systems can help determine vulnerability to land degradation. Such information is important for countries in arid areas with high susceptibility to desertification. SOC stocks, and predicted changes between 2000 and 2030, were determined at the national scale for Jordan using The Global Environment Facility Soil Organic Carbon (GEFSOC) Modelling System. For the purpose of this study, Jordan was divided into three natural regions (The Jordan Valley, the Uplands and the Badia) and three developmental regions (North, Middle and South). Based on this division, Jordan was divided into five zones (based on the dominant land use): the Jordan Valley, the North Uplands, the Middle Uplands, the South Uplands and the Badia. This information was merged using GIS, along with a map of rainfall isohyets, to produce a map with 498 polygons. Each of these was given a unique ID, a land management unit identifier and was characterized in terms of its dominant soil type. Historical land use data, current land use and future land use change scenarios were also assembled, forming major inputs of the modelling system. The GEFSOC Modelling System was then run to produce C stocks in Jordan for the years 1990, 2000 and 2030. The results were compared with conventional methods of estimating carbon stocks, such as the mapping based SOTER method. The results of these comparisons showed that the model runs are acceptable, taking into consideration the limited availability of long-term experimental soil data that can be used to validate them. The main findings of this research show that between 2000 and 2030, SOC may increase in heavily used areas under irrigation and will likely decrease in grazed rangelands that cover most of Jordan giving an overall decrease in total SOC over time if the land is indeed used under the estimated forms of land use. # 2007 Elsevier B.V. All rights reserved. Keywords: Jordan; Century; Roth-C; Soil organic carbon; Soil; Modelling

1. Introduction Regional and global C budget quantifications need to include an understanding of SOC dynamics and SOC distribution at a regional level (Paustian et al., 1997). C sequestration can be indirectly assessed through the * Corresponding author. Tel.: +962 777227472; fax: +962 64871232. E-mail address: [email protected] (R. Al-Adamat). 0167-8809/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.agee.2007.01.006

modelling of SOC content, which may complement direct measurements (Ardo¨ and Olsson, 2003). Modelling helps in identifying areas with large potential for C sequestration. It also helps in predicting and understanding future changes due to climate change, land use change and different land management scenarios (Ardo¨ and Olsson, 2003). Roth-C (Jenkinson and Rayner, 1977) and Century (Parton et al., 1987) are the most widely used SOC simulation models. They have been tested against a variety

36

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

of long-term agricultural field trials in a variety of climate zones, including arid and semi-arid regions. Roth-C requires fewer data inputs than Century and is, therefore, easier to parameterize. However, Roth-C only models soil processes and, consequently, plant residue C is a required input. Century is an ecosystem model that simulates biogeochemical fluxes of C, N, P and S, primary production and water balance on a monthly time step (Parton et al., 1988). The model supplies a tool for ecosystem analysis that enables an evaluation of changes in climate and the management of ecosystems (Ardo¨ and Olsson, 2003). Both models have been used in many parts of the world as tools to predict C stocks and changes (Hill, 2003; Falloon and Smith, 2002; Ardo¨ and Olsson, 2003; Jenkinson et al., 1999). Past studies have used different approaches to integrate Century and Roth-C with spatially explicit databases via geographical information systems (GIS). Fallon et al. (1998) integrated the Roth-C model with GIS to illustrate the effect on SOM through an afforestation scenario in Hungary. Ardo¨ and Olsson (2003) integrated GIS with the Century model to assess SOC in semi-arid Sudan. Refined estimates of potential SOC sources and sinks, including their variation in space and time, are possible through the linkage of dynamic simulation models and spatially explicit data (Ardo¨ and Olsson, 2003). Lal (2002) emphasized that any assessment of soil C at different scales requires GIS and modelling. This paper presents predicted SOC stocks and stock changes (for the years 2000–2030) made for Jordan at the national scale using the GEFSOC Modelling System. This newly developed system links spatially explicit data with two soil C models (Century and Roth-C) and an empirical method for estimating SOC stock changes (the IPCC

method) in a GIS environment. Further details of the development and utilisation of the system are given by Easter et al. (2007).

2. Background information on the study area The study area comprised the whole of Jordan. Jordan is a relatively small country (89,342 km2) located in the eastern Mediterranean region between 298–328N latitude and 348– 398E longitude (Fig. 1). The country is bordered on the north by Syria, to the east by Iraq and by Saudi Arabia on the east and south. To the west is Israel and the occupied West Bank, while Jordan’s only outlet to the sea, the Gulf of Aqaba, is to the south. However, Jordan’s diverse terrain and landscape belie its actual size, demonstrating a variety of landscapes and agro-ecologies, usually found only in large countries. For the purpose of this study, Jordan has been divided into three main geographic and climatic areas: 1. The Jordan Valley, which extends down the entire western flank of Jordan, is the country’s most distinctive natural feature. The Jordan Valley forms part of the Great Rift Valley of Africa, which extends from southern Turkey through Lebanon and Syria to the salty depression of the Dead Sea, where it continues south through Aqaba and the Red Sea to eastern Africa. This fissure was created 20 million years ago by shifting tectonic plates. 2. The Uplands, which separate the Jordan Valley and its margins from the plains of the eastern desert. This region extends the entire length of the western part of the country. The highlands of Jordan host most of Jordan’s

Fig. 1. The location of Jordan in relation to neighbouring countries and rainfall distribution.

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

37

main population centres, including Irbid, Amman, Zarqa, Karak, Tafilah and Shubak. 3. The eastern desert (Badia region), which comprises around 90% of the land area of Jordan. This area of desert and desert steppe is part of what is known as the North Arab Desert. It stretches into Syria, Iraq and Saudi Arabia, with elevations varying between 600 and 900 m above sea level.

3. Methods

2.1. Climate

3.1. Data collection

Jordan is mainly arid to semi-arid and is characterized by dry hot summers and mild winters with extreme variability in rainfall within and among years. More than 90% of the country’s land receives less than 200 mm annual rainfall. This restricts agricultural development and results in a need to import food (Al-Ansari and Baban, 2001). The rain-fed area in Jordan (>200 mm year 1) is limited and restricted to the western and northern highlands (Fig. 1). Rain-fed arable lands have been extensively degraded due to population pressure and climate and to the struggle to produce more in order to satisfy increasing food demands. SOM is generally low in arid and semi-arid areas and is exacerbated by overgrazing of crop residues resulting in very low levels of SOM and increasing the potential of water and wind erosion of the topsoil. Most of the rain falls between November and March. Rainfall decreases from west to east and from north to south. The Rainfall in Jordan is characterized by irregularity and high variability with a spatial distribution closely related to topography. Table 1 shows four broad climatic zones for the country, defined in terms of isohyets of 100 mm year 1, and how these zones are distributed over the three main geographic regions of Jordan. For example, in view of the varied topography, a fairly wide range in annual rainfall is observed within the Uplands.

In order to make estimates of SOC content, SOC stocks and changes for Jordan using the GEFSOC Modelling System, ideally two types of data are needed: national scale data sets (soils, climate, land use and land history) to act as model input data and site-scale data for model evaluation (further details of the stages involved in using the GEFSOC Modelling System are given in Milne et al., 2007 and Easter et al., 2007). Ideally the two soil C models (Roth-C and Century) in the GEFSOC System should be validated using long-term experimental datasets from the region in question (Easter et al., 2007). However, following an extensive study of available information from a wide variety of sources, it was found that for Jordan, no such data sets were available.

2.2. Land use Jordan has a long history of agricultural land use. Irrigation in Jordan has been practiced since pre-Roman times. For centuries there has been sufficient vegetative cover to meet the domestic needs for wood and to provide grazing for a large number of animals. Among the crops, which were extensively grown during ancient times and still occupy a considerable proportion of the dry lands, are Table 1 Climatic zones of Jordan 2

Area (km )

Badia Uplands Jordan Valley Total

Rainfall (mm) 400

Total

69,101 5,080 2,478 76,560

2816 3420 651 6885

197 3496 300 3993

0 1806 0 1806

72,015 13,801 3,428 89,342

Source: Ministry of Agriculture, The Government of Jordan, 2004.

cereals (e.g. Triticum aestivum L., Hordeum vulgare), grapes (Vitis vinifera), and olives (Olea europaea L.). Land use in Jordan falls into four broad types that reflect topography, climate, especially rainfall, and the availability of supplementary moisture supply as shown in Table 2.

3.2. National scale GEFSOC System input data 3.2.1. Climate data The climate data were acquired from the Jordan Meteorological Department (JMD). For some stations, these data go back to the 1960s. The data included evaporation, air temperature and precipitation. These data came from 30 meteorological stations distributed throughout Jordan. 3.2.2. Crop statistics The Department of Statistics (DOS) of The Jordanian Government has a large database of crops that are grown in Jordan. The main parameter needed for this research was crop yield for vegetables (e.g. Allium sp., Lycopersicon esculentum L.), bananas (Musa  paradisiaca L.), citrus (Citrus sp.), cereals and fruit trees for each ecological zone in Jordan as shown in Table 3. These data also allowed the adaptation of existing, or the creation of new crop files (that are specific to Jordan) within the Century model. Table 2 Broad types of land use in Jordan Land zone

Description

Rain-fed agricultural

Tree crops within the hilly and steeply sloping lands and wheat (Triticum aestivum) on the undulating lands Steppe grassland and brush species Rangelands The Jordan Valley: citrus (Citrus sp.), banana (Musa  paradisiaca L.), vegetables, fruits and cereals

Low rainfall The Desert Irrigated agriculture

38

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

Table 3 Crop yield for a selection of crops in Jordan (t ha 1) Region

Vegetables

Citrus (Citrus sp.)

Banana (Musa  paradisiaca L.)

Fruits

Cereals

Olives (Olea europaea)

Jordan Valley Badia Uplands

17.0 29.0 –

10.0 – –

10.0 – –

15.0 – 5.0

9.0 2.8 9.0

3.5 3.0 –

Source: Department of Statistics, The Government of Jordan, 2004.

3.2.3. Soil data Soil data were derived from a 1:500,000 scale Soil and Terrain (SOTER) database (NSMLUP, 1996; Batjes et al., 2003, 2007), compiled from the 1:250,000 generalized soil map of Jordan (MINAG, 1993). The SOTER database includes information on SOC, soil texture and bulk density for 28 soil units in a GIS format. 3.3. Validation data To the authors knowledge, there are no long-term experimental data sets in Jordan that have complete measurements of all the parameters necessary to evaluate the performance of Roth-C and Century for Jordanian conditions. The National Centre for Agricultural Research and Technology Transfer (NCARTT) has seven experimental stations, which have a long history of land use investigations and these data were collated as part of this study. However, the data collected from NCARTT has several problems that prevented it from being used for model validation, e.g.: 1. Lack of coordinates that could be used to spatially join different measurements of SOM over time. 2. The lack of supporting information such as the type of management employed over a specific area in terms of crop rotation, fertilizer inputs, soil texture and irrigation. Data were also available from ICARDA for the Aleppo region of Syria. However, this particular region of Syria was not considered to be representative of any region in Jordan in terms of soil texture and climatic parameters. The ideal of being able to validate Roth-C and Century using site-scale long-term experimental data was, therefore, not possible. National scale SOC stock estimates made using the GEFSOC System for the year 1990 for Jordan were compared with estimates made using mapping based approaches, in order to see if estimates were reasonably similar (see Section 3.6). 3.4. Data preparation 3.4.1. GEFSOC System input data The available climatic, land use and soil data were formatted to enable input to the GEFSOC Modelling System; the procedure is detailed in Easter et al. (2005).

3.4.2. Land use scenario development SOC stocks are dependant, to a large extent, on land use history, with changes in land use having potential effects on SOC content for decades and even centuries (Pulleman et al., 2000). To predict current and future soil C stocks, historical land use, current land use and land use change scenarios had to be devised for the whole of Jordan. Historical and current land use was based on records of the Jordanian Government’s Department of Statistics (http://www.dos.gov.jo) and expert opinion. Based on the statistical data and FAO (2003) projections for 2015 and 2030 and to achieve the objectives of this research project, the Uplands region was divided into three zones based on the dominant land use in order to facilitate the modelling procedure, namely the North Uplands, Middle and South Uplands. This was achieved by intersecting the developmental regions map and the Jordan Ecological Zones map (RJGC, 1995) using GIS. Thus Jordan was divided into five zones: the Badia, the North Uplands, the Middle Uplands, the South Uplands and the Jordan Valley. The scenarios for current land management used in this project were based on the records of the Department of Statistics (http://www.dos. gov.jo). Table 4 provides a summary of the major land use, considered in this research, in all five zones of Jordan. 3.5. Estimation of SOC stocks and changes using the GEFSOC Modelling System After compilation and formatting of national scale datasets of soils, climate and land use for Jordan, the GEFSOC Modelling System was run to produce estimates of SOC stocks and changes from the three methods included in the system (the Roth-C and Century models and the IPCC method). The system was run through an equilibrium period (approximately 10,000 years) through historical, current and finally future land use. The system generates large amounts of data. To reduce the size of the output dataset, whilst still producing meaningful results for users, the modelling system generates a series of regression statistics based upon user-defined breakpoints in the management sequence (Easter et al., 2007). In this case the years 1990, 2000 and 2030 were specified. 3.6. Comparison with SOTER-derived estimates Batjes et al. (2007) used the Jordanian SOTER database to estimate SOC stocks using a system of

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

39

Table 4 Historical land use and future land use scenarios for Jordan Region period

North Uplands

Middle Uplands

South Uplands

Badia

Jordan Valley

Pre-1900 1950

Short grass and forest Short grass, olives (Olea europaea), fruits, cereals and forest Short grass, olives, cereals, fruits, forest and urban Short grass, olives, cereals, fruits, forest and urban Short grass, olives, cereals, fruits, forest and urban

Grass Grass, cereals and urban

Grass Grass and cereals

Rangelands Rangelands

Savannah Cereals

Grass, cereals, olive and urban

Grass, cereals, olives, fruits and urban

Rangelands, vegetables, cereals and abandoned lands

Grass, cereals and urban

Grass, cereals, olives, fruits and urban

Rangelands, vegetables, cereals and abandoned lands

Citrus (Citrus sp.), banana (Musa paradisiaca L.), vegetables, fruits, cereals and urban Citrus, banana, vegetables, fruits, cereals and urban

Grass, cereals and urban

Grass, cereals, olives, fruits and urban

Rangelands, vegetables, cereals and abandoned lands

Citrus, banana, vegetables, fruits, cereals and urban

2000

2015

2030

taxotransfer rules and the simulation of phenoforms; this method allowed them to define 95%-confidence intervals for median soil C stocks at a regional scale. In addition to this, data were also generated using three other methods (Batjes, 2004), two of which are considered here. With SOTER method A, the soil C content in the 0–30 cm soil layer is computed by SOTER unit, using data from the corresponding representative soil profiles; results are then linked to the spatial information using GIS. SOTER method B is similar to method A, but uses the average content of soil C computed per FAO soil unit. In this paper, stocks derived using methods A and B were compared with the C stocks for 1990 estimated using the Century and Roth-C modules of the GEFSOC Modelling System.

4. Results

for the model runs. In this study, results were taken for the years 1990 (to represent the Kyoto Protocol baseline year), 2000 (to represent the ‘current’ situation) and 2015 and 2030 to represent the future. 4.1.1. Model estimate of current regional C stocks— Century output from the GEFSOC System Table 5 shows national SOC stocks for Jordan made using the three methods in the GEFSOC Modelling System. The total national SOC stock in 1990 was 71 Tg according to Century output. As expected, national stocks were low due to the small area covered by the country but also because of the arid climate in Jordan. Fig. 2 shows C content (t C ha 1) in 1990 for Jordan, using the Century output from the GEFSOC Modelling System. This figure shows that in 1990, the maximum organic C content (0–20 cm soil depth) was in the Middle Uplands region (500 mm of rainfall) and was 27 t ha 1. In the Badia region, the C content varied from

We used the GEFSOC Modelling System to make spatially explicit estimates of SOC stocks and changes based on three different methods—two modelling approaches (Century and Roth-C) and the empirical IPCC method (see Easter et al., 2007 for details). 4.1. SOC stocks and changes using the GEFSOC Modelling System: current Results of the GEFSOC Modelling System can be produced for any given year during the total period chosen Table 5 National SOC stock estimates for Jordan made using the GEFSOC Modelling System Year

Model/method Roth-C (0–20 cm)

Jordanian 1990 2000 2030

Century (0–20 cm)

National SOC Stock (Tg) 104 71 102 66 100 57

IPCC (0–30 cm) 237 242 249

Fig. 2. Estimates of regional soil organic carbon content in 1990 using Century output from the GEFSOC Modelling System.

40

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

Fig. 3. Estimates of regional soil organic carbon content in 2000 using Century output from the GEFSOC Modelling System.

2.5 to 17.4 t ha 1 (increasing with rainfall). In the North, Middle and South Uplands the ranges were 3.6–20 t ha 1, 4.4–26 t ha 1 and 2.8–20.9 t ha 1, respectively, also increasing with rainfall. In the Jordan Valley, C content ranged from 16.2 to 25 t ha 1. The high SOC in some parts of the Badia may be due to the fact that in 1990 many areas of rangeland had not yet been severely affected by overgrazing. The Jordan Valley has a very complex farming system with heavy use of fertilizers and manures. According to the Century output, national SOC stocks for the year 2000 declined in relation to 1990 levels with a stock of 66 Tg in 2000 (Table 5). Fig. 3 shows SOC content in Jordan in 2000 using the Century output from the GEFSOC Modelling System. It shows that the maximum C content in the upper 20 cm was estimated at less than 27 t ha 1. From a regional perspective, the Badia had 2.3–15.3 t C ha 1 (linear increase with rainfall) and the North Uplands had 4.1– 25 t C ha 1. The Middle Uplands had 4.4–23.9 t C ha 1. In the South Uplands, C content ranged from 2.8 to 19 t ha 1. The Jordan Valley had 17.6–26.8 t C ha 1. These numbers showed that there was a decrease in SOC in the Badia between 1990 and 2000 that could be explained by the overgrazing that occurred during the 1990s. In the uplands, SOC increased in the North, the middle and the South due to the introduction of olives and fruit trees and the continuation of afforestation. In the Jordan Valley, an increase of SOC also occurred between 1990 and 2000 due to an increase in the area under banana and citrus. 4.1.2. Model estimate of current regional C stocks— Roth-C output from the GEFSOC System The Roth-C estimates from the GEFSOC System for 1990 were on average higher than those given by the Century

Fig. 4. Estimates of regional soil organic carbon content in 1990 using Roth-C output from the GEFSOC Modelling System.

output. Total national stock was 104 Tg (Table 5). The difference could possibly be due to the different ways the two models deal with the effects of soil moisture on decomposition. Despite the difference in total stocks, the relative distribution of soil C stocks among the different land uses were similar to that predicted by Century. SOC was highest (reaching nearly 70 t ha 1) in the Jordan Valley (Fig. 4) where the land use is most intensive and where C addition rates are highest due to the existence of many types of crops (including tree crops) and the extensive use of irrigation water, chemical fertilizers and animal manures. In the Badia, SOC varied between 4 t ha 1 in lower rainfall zones and 27.8 t ha 1 in higher rainfall areas, which is 31% higher than Century estimates for the same region. In the North Uplands, SOC ranged from 10.9 to 36.4 t ha 1, ca. 55% higher than Century estimates. In the Middle Uplands, an SOC content of 6.7–34.9 t ha 1 was estimated using Roth-C, 45% higher than the Century estimate. The South Uplands SOC estimates varied between 4.3 and 33.4 t ha 1, 50% higher than Century estimates. In the Jordan Valley, Roth-C estimates of SOC were very high when compared with Century. SOC estimates ranged from 57.9 to 69 t ha 1 ca. 68% higher than Century estimates. According to Roth-C, national SOC stocks for Jordan totalled 102 Tg in 2000 (Table 5). The upland areas in Jordan had SOC content of 11.6–36.7 t ha 1 in the North, 6.5– 33.4 t ha 1 in the Middle and 4.3–33.4 t ha 1 in the South (Fig. 5). The Badia SOC estimates varied between 4.0 and 27.7 t ha 1 and the Jordan Valley between 67 and 80 t ha 1. These estimates for all areas in Jordan were higher than Century estimates by 35% for the Badia, 53% for the North Uplands, 46% for the Middle Uplands, 48% for the South Uplands and 70% for the Jordan Valley.

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

41

Fig. 5. Estimates of regional soil organic carbon content in 2000 using Roth-C output from the GEFSOC Modelling System.

Fig. 6. Estimates of regional soil organic carbon content in 2030 using Century output from the GEFSOC Modelling System.

As with Century, Roth-C estimates for 1990 and 2000 had a linear relationship with rainfall in the Badia and the Uplands. Roth-C estimates were higher than Century estimates by approximately 45%. Total current SOC stock estimated by the IPCC method for the whole of Jordan was 237 and 240 Tg for 1990 and 2000, respectively, much higher than the estimates made using Century and Roth-C (Table 5). This is undoubtedly an overestimate and is due to the fact that the default (Tier 1) reference C stock values in the IPCC method were not appropriate for a large part of Jordan.

28 and 29 t ha 1, respectively. The highest SOC values in Jordan (specifically the Jordan Valley) in 2015 and 2030 according to the Roth-C output reach a maximum of 89 and 99 t ha 1 in 2015 and 2030, respectively. Table 6 shows SOC estimates using the Century and Roth-C models for 1990, 2000, 2015 and 2030 by region based on the land use management scenarios used in this study. Century predicts a substantial decrease in SOC content in the Badia region from 6.6 t C ha 1 in 2000 to 5 t C ha 1 in 2030 but Roth-C predicts only a small decrease. When dealing with the hyperarid environment of the Badia, this is again probably due to the different ways the two models deal with the effects of soil moisture on decomposition. In the North Uplands, there will be a projected increase between 2000 and 2015 and then SOC is projected to decrease in 2030, which can be attributed to an expected increase in urbanization (1% annually). Although urbanization may not necessarily reduce SOC content, in this study, the assumption was made that soils under urbanized land have no capacity for SOC increase. In the Middle Uplands, a continuous decrease in SOC is predicted by both models which is also a in part a result of assuming more urbanization occurring in the region due to the continuous expansion of

4.2. SOC stocks and changes using the GEFSOC Modelling System: future Century, Roth-C and IPCC estimates of National SOC stock for Jordan in 2030 were 57, 100 and 249 Tg, respectively. Again estimates made using the IPCC method were higher than expected and Roth-C estimates were roughly twice those made by Century. Fig. 6 shows the estimated regional SOC content in Jordan in 2030 using the Century output. The maximum estimated SOC content in Jordan in 2015 and 2030 according to Century was less than

Table 6 Average estimates of soil organic carbon content (t C ha 1) in 1990, 2000, 2015 and 2030 using the Century and Roth-C models Region

Badia North Uplands Middle Uplands South Uplands Jordan Valley

Century

Roth-C

1990

2000

2015

2030

1990

2000

2015

2030

7.2 7.3 7.5 7.4 20.0

6.8 7.9 7.2 7.1 21.9

5.9 8.4 7.0 6.8 22.9

5.0 7.8 6.7 6.4 24.0

11.4 15.8 13.5 15.0 62.0

11.5 16.5 13.4 13.7 72.5

11.4 15.8 13.1 13.7 81.1

11.2 17.1 12.8 13.0 90.8

Note. All values are for 0–20 cm depth.

42

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

Amman, Salt, Madaba and Zarqa (ca. 80% of Jordan’s population). The South Uplands would also experience a decrease in SOC based on the land use management scenarios for the region, which also assumed increased urbanization. Only the Jordan Valley would have an increase in SOC content in 2015 and 2030 resulting from vegetables being replaced by banana and citrus. The predicted increase in the Jordan Valley using Roth-C, however, is higher than would be expected.

5. Discussion 5.1. SOC estimates from the three GEFSOC methods As pointed out in Section 4, SOC stock estimates produced by the IPCC computational method, were much higher than those produced by the modelling methods. This overestimate is attributable to the use of global default values. The IPCC default reference C stock levels for the ‘warm-temperate dry’ region range from 19 to 38 Mg C ha 1 (depending on soil type); thus even after applying appropriate reduction factors to reflect degradation and other (negative) effects of management, mean SOC stocks cannot be less than about 50% of the reference values using the IPCC method. This highlights a limitation of using global defaults in the IPCC model, where use of countryspecific data for reference C stocks would considerably improve accuracy of the method. Roth-C estimates of SOC stocks were 45% higher than estimates made using Century. This discrepancy is not entirely unexpected as neither of the models was developed specifically for use in hyper-arid conditions such as those found in Jordan. As mentioned in Section 4 this could be due to the different ways the models account for the effects of soil moisture on decomposition rates. The GEFSOC Modelling System uses Roth-C 26.3, as this is the publicly available version of the model. Jenkinson et al. (1999) used a modified version of the Roth-C model (Roth-C 26.5) to simulate organic matter turnover in a long-term experiment in Syria. In this modified version soils are allowed to more completely dry out than in Roth-C 26.3 and modifications have been made to decomposition rates once the soil reaches wilting point. It could be that the Roth-C overestimations of SOC we found for Jordan are due to the models inability to simulate soil drying and subsequent effects on decomposition rates in the hyper-arid conditions found in the Jordan Badia (which covers 80% of the area of the country). Century deals with soil moisture differently from Roth-C as it includes a water budget sub-model that simulates movement of water through soil layers, in addition to loss from evaporation and transpiration. The difference between the two models in terms of the way they deal with soil moisture and its impact on decomposition rates in the very arid conditions found in Jordan is something that needs further investigation. Lack of suitable long-term experi-

mental datasets from Jordan limited our ability to do this before national scale model runs were carried out. 5.2. Comparison of GEFSOC SOC estimates with output from SOTER methods In this research, as addressed earlier, there were no data available to validate the models in order to make estimates of C stocks in Jordan more specific to the conditions found in the country. However, it was possible to compare the resulting national estimates of soil C stocks for 1990 with estimates made using SOTER mapping based methods for the same year. SOTER methods calculate the C stocks for the top 30 cm, while the Century estimates are for the top 20 cm. In order to make both estimates for the same depth, it was assumed that all soil parameters have a linear relationship with depth. This assumption meant that all SOTER values where multiplied by 2/3. As a result, the maximum C stock in Jordan – computed for xeric regions in the northern highlands – using the SOTER A method is approximately 23 t ha 1 in the top 20 cm of soil. Using the SOTER B method, the maximum C stocks in Jordan are 18 t ha 1 for the top 20 cm. In Table 7 a comparison is made between C stocks estimate for 1990 made using SOTER methods A and B and Century and Roth-C. It appears from this table that the results of both Century and Roth-C are in the same range given the data limitation addressed earlier in this paper and the fact that SOC is generally higher in the top 20 cm of soil and does not, in fact, have a linear relationship with depth. In the Jordan Valley, Roth-C estimates are high compared with SOTER A and B and Century estimates are slightly higher but within the same range. The differences between estimates made using SOTER methods and Century and Roth-C estimates could be due to the following reasons: SOTER data are based on available (historic) soil profile data and as such do not account explicitly for changes in land use over time. Data for the SOTER for Jordan were collected between 1990 and 1992 (MINAG, 1993). Conversely, the GEFSOC modelling system estimates are based on historical land use data that go beyond 1900, land use management scenarios, crop yield and climatic factors that affect the change in C stocks in the soil. SOTER methods estimate the C stocks for the top 30 cm of soil while Century and Roth-C estimates are for the top 20 cm of soil; the assumption of a linear decrease in SOC with depth, used here for comparing SOTER-based with GEFSOC-modelled results, however, may be invalid for soils of hyper-arid and arid regions, where most SOC is often concentrated in the upper few centimeters of the soil profile (see NSMLUP, 1996). Limited field data to validate Century and Roth-C models, and hence limited scope for detailed evaluation of the modelled output. Finally, the GEFSOC Modelling System only considers regional differences in soil texture and soil wetness, whereas

Average

13.8 12.1 20.0 62.0

Maximum

22.7 12.7 25.2 69.1

regional differences in other soil properties that affect SOC stocks and change are implicitly considered in the SOTER methods. For example, soil salinity, sodicity and presence of coarse fragments (>2 mm), all of which are of common occurrence in large sections of Jordan (NSMLUP, 1996).

10 8.7 16.2 57.9 10.6 7.0 7.4 15 22.7 10.7 20.9 33.4 2.0 3.3 2.8 4.3 16.0 10.4 7.5 13.5 22.7 18.0 26.0 34.9 12 6.0 4.4 6.8 14.4 10.2 7.3 15.8 22.7 18 20.7 36.4 3.3 4.0 3.7 10.9 Note. All values are for 0–20 cm depth.

6.3 6.3 7.2 11.4 22.7 18.0 17.5 27.8 1.3 3.0 2.5 4.0

Maximum

43

5.3. SOC stocks and land use in Jordan

SOTER A SOTER B Century Roth-C

Minimum Average Maximum

South Uplands

Minimum Average Maximum

Middle Uplands

Minimum Average Maximum

North Uplands

Minimum Minimum

Average Badia Method

Table 7 Comparison of SOC contents (t C ha 1) produced using SOTER methods and Century and Roth-C output from the GEFSOC Modelling System (for the year 1990)

Jordan Valley

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

When the country is considered as a whole the Badia region makes the largest contribution to total SOC stock due to its large area (the Badia covers approximately 80% of Jordanian territory). It also has the lowest SOC content in t C ha 1 as most of the area comprises of sparse rangeland that receives less than 150 mm year 1 rainfall (Findlay and Maani, 1998). The national decline in SOC stock predicted by both Roth-C and Century are due, almost exclusively, to the degradation of rangeland and pasture land in the Badia. Excessive overgrazing began in the Badia during the 1990s when there was a massive increase in the number of sheep and goats in the region, producing a legacy effect on SOC stocks in the year 2000 and beyond. In this study a continuation of overgrazing leading to a conversion of remaining good rangeland into degraded rangeland or abandoned areas was assumed for 2000–2030. This was based on expert opinion and extrapolation of existing trends. If this scenario is correct, action is needed by land use planners in Jordan if severe SOC losses and further land degradation in the area are to be avoided. Drivers of land use change in Jordan are extremely complicated and have changed markedly over the last century. Until the second half of the 20th century the population of the area that is now Jordan were rural and population levels numbers were dictated by the environmental resources available. However, from the mid-20th century onwards, population was also driven very much by migration from neighbouring countries facing political instability. Urbanization increased and traditional land management strategies, such as nomadic pastoralism, decreased. In addition, Jordan still has one of the highest rates of natural population increase in the world (Dutton, 1998). Predicting population change and its impact on future land use and subsequently SOC stocks is, therefore, difficult. This study predicted some increase in SOC under intensive cropland (which is mainly confined to the Jordan valley). Increases are due to projected improvements in production and C inputs. From a national perspective the area concerned is small and thus impacts on total SOC stocks are minimal. However, from a land degradation point of view, this could be an important estimation, given the disproportionate number of Jordanians who live in the Jordan valley and therefore rely on the continued productivity of this area. It should be pointed out that the GEFSOC Modelling System does not yet account for possible salinisation effects associated with intensive irrigation practices, such as those in the Jordan valley and any SOC increases in this area could, therefore be

44

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45

overestimated. It should also be noted that this study used current weather data and did not model the potential impacts of predicted climate change in Jordan. If climate were to become hotter and drier in the region with more erratic rainfall patterns this could decrease the amount of water available for irrigation leading to a reduction in productivity and a subsequent reduction in SOC. 5.4. SOC stock change estimates in Jordan and UN reporting Jordan is a signatory country to The United Nations Framework Convention on Climate Change. As such it is obliged to report estimated C emissions to the atmosphere from all sources, including those resulting from land use change. The estimates made in this study are the only SOC stock change estimates that have been made for Jordan using an IPCC (Intergovernmental Panel on Climate Change) ‘Tier 3’ advanced inventory method. It is hoped that as the GEFSOC Modelling System is further developed and improved, it will be increasingly used by countries in arid and semi-arid areas in the production of national GHG inventories. This application will be timely, given the current interest in/debate about the potential of degraded lands in arid areas for C sequestration under appropriate land management.

problems associated with intensive irrigation are advised to improve SOC stock estimates for the Jordan Valley. In conclusion, this research project has provided a great opportunity to collect and collate various types of data in one location for the first time in Jordan. It also, provided acceptable SOC stock estimates in a country where data is limited and has provided tools that can be improved in the future when more information becomes available. It is recommended that in future projects of this kind, more soil sampling is undertaken to allow validation of the models, which will eventually lead to improved estimates of SOC stocks. The results of this research project could help decision makers in Jordan make better land use/management plans in the future, especially in the Badia region where rangelands are under continuous degradation. More rangeland reserves in the Jordanian Badia and better management of the limited vegetation cover will eventually improve C sequestration in this vast area. In return, this will help Jordan implement the Kyoto protocol in terms of improved C sequestration and it will also help reduce national CO2 emissions and therefore assist in tackling global warming. Finally, it is recommended that a simple long-term experiment be established in Jordan, either by the Jordan Ministry of Agriculture or by Jordan Badia Research and Development Centre. This would provide much needed data for a range of land use conditions providing a calibration data set to further refine future model predictions.

6. Conclusions Based on the land use management scenarios suggested in this research project and the Century output of the GEFSOC Modelling System, a decrease in the C stocks in the Badia is expected in 2030 compared with 1990. Also, in the Northern Uplands, there would be an increase in the C stocks in 2015 compared with 2000 and a decrease in 2030, which could be due to a projected increase in urbanization. In the Middle Uplands, a continuous decrease in C stocks is predicted by both models. The South Uplands are also expected to have a decrease in C stocks based on the land use management scenarios for the region, which again assumes more urbanization. Only, the Jordan Valley will have more C stocks in 2015 and 2030 because of an increase in citrus and banana trees at the expense of vegetables. In general, it was found that there is a linear relationship between rainfall and SOC in Jordan. The Jordan Valley is an exception due its complexity and the use of irrigation water. The Century and Roth-C estimates of SOC made using the GEFSOC System were compared with output from mapping based approaches. Based on this comparison, and taking into consideration the limitation addressed earlier in this paper, results of both Century and Roth-C are in the same range for most of Jordan except for the Jordan Valley, where Roth-C estimates are high when compared with estimates made using SOTER methods A and B. Improvements to the crop productivity sub-models of Century and modifications in the way that both models treat salinity

Acknowledgments We would like to thank the Badia Research and Development Center for supporting this work. The project Assessment of Soil Organic Carbon Stocks and Changes at National Scale was co-financed by the GEF (GFL-2740-024381), implemented by UNEP, and coordinated by the University of Reading, UK. It was carried out by a consortium of partners from Austria, Brazil, France, India, Jordan, Kenya, the Netherlands, the United Kingdom and the USA with supplemental funding from a wide range of sponsors (see http://www.nrel.colostate.edu/projects/gefsoc-uk for details). Several national and international agencies were contacted to collect the necessary data for this research, including; Royal Jordanian Geographic Centre (RJGC), United Nations University (UNU), Ministry of Agriculture (MOA), University of Jordan, (UJ), International Centre for Agricultural Research in the Dry Areas (ICARDA), Department of Statistics (DOS) and The Statistical, Economic and Social Research and Training Centre for Islamic Countries (SESRTCIC).

References Al-Ansari, N.A., Baban, S.M.J., 2001. The climate and water resources. In: Baban, S.M.J., Al-Ansari, N.A. (Eds.), Living with Water Scarcity:

R. Al-Adamat et al. / Agriculture, Ecosystems and Environment 122 (2007) 35–45 Water Resources in Jordan, Badia Region, The way forward. Al al-Bayt University, Jordan. Ardo¨, J., Olsson, L., 2003. Assessment of soil organic carbon in semi-arid Sudan using GIS and the CENTURY model. J. Arid Environ. 54, 633–651. Batjes, N.H., 2004. Soil carbon stocks and projected changes according to land use and management: a case study for Kenya. Soil Use Manage. 20, 350–356. Batjes, N.H., Rawajfih, Z., Al-Adamat, R., 2003. Soil data derived from SOTER for studies of carbon stocks and change in Jordan, Version (1.0; GEFSOC Project;), Technical Report 2003/04. ISRIC - World Soil Information, Wageningen, The Netherlands. Batjes, N.H., Al-Adamat, R., Bhattacharyya, T., Bernoux, M., Cerri, C.E.P., Gicheru, P., Kamoni, P., Milne, E., Pal, D.K., Rawajfih, Z., 2007. Preparation of consistent soil data sets for modelling purposes: secondary SOTER data for four case study areas. In: Milne, E., Powlson, D.S., Cerri, C.E.P. (Eds.), Soil Carbon Stocks at Regional Scales. Agric. Ecosyst. Environ. 122, 26–34. Dutton, R., 1998. The Jordan Badia programme: population–environment interactions. In: Clarke, J., Nion, D. (Eds.), Population and Environment in Arid Regions. Parthenon Publishing Group Inc., New York, p. 232. Easter, M.J., Paustian, K., Killian, K., Boyack, T., Williams, S.A., Feng, T., Coleman, K., Swan, A., Al-Adamat, R., Bhattacharrya, T., Cerri, C.E.P., Kamoni, P., Batjes, N.H., Milne, E., 2005. User Manual: GEFSOC Soil Carbon Modeling System. Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, CO. , http://www.nrel.colostate. edu/projects/gefsoc/download.htm. Easter, M., Paustian, K., Killian, K., Williams, S., Feng, T., Al Adamat, R., Batjes, N.H., Bernoux, M., Bhattacharyya, T., Cerri, C.C., Cerri, C.E.P., Coleman, K., Falloon, P., Feller, C., Gicheru, P., Kamoni, P., Milne, E., Pal, D.K., Powlson, D.S., Rawajfih, Z., Sessay, M., Wokabi, S., 2007. The GEFSOC soil carbon modelling system: a tool for conducting regionalscale soil carbon inventories and assessing the impacts of land use change on soil carbon. In: Milne, E., Powlson, D.S., Cerri, C.E.P. (Eds.), Soil Carbon Stocks at Regional Scales. Agric. Ecosyst. Environ. 122, 13–25. Fallon, P.D., Smith, P., Smith, J.U., Szabo, J., Coleman, K., Marshall, S., 1998. Regional estimates of carbon sequestration potential: linking the Rothamsted Carbon Model to GIS databases. Biol. Fert. Soils 27, 236–241. Falloon, P., Smith, P., 2002. Simulating SOC changes in long-term experiments with RothC and CENTURY: model evaluation for a regional scale application. Soil Use Manage. 18, 101–111. FAO (Food and Agriculture Organization), 2003. World Agriculture: Towards 2015/2030. Food and Agriculture Organization of the United Nations, Rome, 97 pp. Findlay, A.M., Maani, M., 1998. Fertility trends in an arid environment: the population of the Jordanian Badia. In: Clarke, J., Nion, D. (Eds.),

45

Population and Environment in Arid Regions. Parthenon Publishing Group Inc., New York, p. 232. Hill, M.J., 2003. Generating generic response signals for scenario calculation of management effects on carbon sequestration in agriculture: approximation of main effects using CENTURY. Environ. Modell. Softw. 18, 899–913. Jenkinson, D.S., Harris, H.C., Ryan, J., McNeill, A.M., Pilbeam, C.J., Coleman, K., 1999. Organic matter turnover in a calcareous clay soil from Syria under a two course cereal rotation. Soil Biol. Biochem. 31, 687–693. Jenkinson, D.S., Rayner, J.H., 1977. The turnover of soil organic matter in some of the Rothamsted classical experiments. Soil Sci. 123, 298–305. Lal, R., 2002. Soil carbon dynamics in cropland and rangeland. Environ. Pollut. 116, 353–362. Milne, E., Al-Adamat, R., Batjes, N.H., Bernoux, M., Bhattacharyya, T., Cerri, C.C., Cerri, C.E.P., Coleman, K., Easter, M., Falloon, P., Feller, C., Gicheru, P., Kamoni, P., Killian, K., Pal, D.K., Paustian, K., Powlson, D., Rawajfih, Z., Sessay, M., Williams, S., Wokabi, S., 2007. National and sub national assessments of soil organic carbon stocks and changes: the GEFSOC modelling system. In: Milne, E., Powlson, D.S., Cerri, C.E.P. (Eds.), Soil Carbon Stocks at Regional Scales. Agric. Ecosyst. Environ. 122, 3–12. MINAG, 1993. National Soil Map and Land Use Project—The Soils of Jordan (Level 1: Reconnaissance Soil Survey; vol.3: Representative Profiles and Soil Analyses) Hunting Technical Services Ltd. in association with Jordanian Soil Survey and Land Research Centre, Ministry of Agriculture, Amman. NSMLUP, 1996. The Soil and Terrain Database for Jordan at scale 1:500,000 (ver. 1. 0; unpublished) National Soil Map and Land Use Project, Department of Afforestation and Forests, Ministry of Agriculture, Amman. DOS (Department of Statistics), 2004. Agriculture survey, online at: http:// www.dos.gov.jo/agr/agr_e/index.htm. Parton, W.J., Schimel, D.S., Cole, C.V., Ojima, D.S., 1987. Analysis of factors controlling soil organic levels of grasslands in the Great Plains. Soil Sci. Soc. Am. J. 51, 1173–1179. Parton, W.J., Stewart, J.W.B., Cole, C.V., 1988. Dynamics of C, N, P and S in grasslands soils: a model. Biogeochemistry 5, 109–131. Paustian, K., Levine, E., Post, W.M., Ryzhova, I.M., 1997. The use of models to integrate information and understanding of soil C at the regional scale. Geoderma 79, 227–260. Pulleman, M.M., Bouma, J., van Essen, E.A., Meijles, E.W., 2000. Soil organic matter content as a function of different land use history. Soil Sci. Soc. Am. J. 64 (2), 689–693. RJGC (Royal Jordanian Geographical Centre), 1995. Jordan Ecological Zones Map. Royal Jordanian Geographical Centre available from http:// www.rjgc.gov.