Assessment of Greenhouse Gas Emissions from Different Land-Use Systems : A Case Study of CO 2 in the Southern Zone of Ghana

Soil and Irrigation Research Centre, College of Basic and Applied Sciences, University of Ghana, Legon, Ghana CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS), International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), BP 320 Bamako, Mali Laboratory of Hydraulics and Environmental Modeling (HydroModE-Lab), Faculty of Agronomy, Université de Parakou, 03 BP 351, Parakou, Benin World Agroforestry Centre (ICRAF), West and Central Africa Regional Office–Sahel Node, BP E5118 Bamako, Mali Department of Soil Science, College of Basic and Applied Sciences, University of Ghana, Legon, Ghana


Introduction
Land-use and land-cover change is among the most important human alterations of the Earth's land surface [1].Conversion or overutilization of land by processes such as cultivation, excessive removal of vegetation, burning, tree plantation, and other forms of degradation and restoration can add or remove greenhouse gases (GHGs) from the atmosphere and thereby impact on the global carbon cycle [2].GHGs are substances believed to make the atmosphere function like the glass in a greenhouse.ey trap the sun's shortwave energy and re-emit it as heat-producing longwave radiation, causing an increase in atmospheric temperature [3].GHG emissions and their interaction with radiation are believed to be the major cause of global climate change, which has become a major threat to development and food security, especially in the tropics [4,5].e anthropogenic gases that are primarily responsible for causing the greenhouse effect include CO 2 , methane (CH 4 ), nitrous oxide (N 2 O), sulphur hexafluoride (SF 6 ), perfluorocarbons (PFCs), and hydrofluorocarbons (HFCs).In 2010, CO 2 , N 2 O, and CH 4 accounted for 66.5%, 17.2%, and 15.4% of greenhouse gases, respectively, worldwide [6].
e level of CO 2 in the atmosphere is estimated to have been ∼280 ppmv on average during the preindustrial period before rising from ∼315 ppmv in 1957 to ∼356 ppmv in 1993 when more accurate monitoring began [7].e current rate of increase of ∼1.5 ppmv/yr is due to the combustion of fossil fuels, cement production, and land-use conversion [3].Agriculture accounts for approximately 10-12% of total global anthropogenic emissions of GHGs, which amounts to 60% and 50% of global N 2 O and CH 4 emissions, respectively [8].In tropical countries, a great amount of the CO 2 emissions steams from vegetation removal, burning and decomposition, and soil carbon loss due to cultivation and soil degradation [9].e CH 4 and N 2 O emissions emanate from marshy fields such as those found in lowland rice production systems, but also from animal production sites.Recent results of a meta-analysis of N fertilizer effects on GHG emissions showed that the N fertilizer-induced N 2 O emission factor during the rice growing season was 0.21% for continuously flooded rice systems and 0.40% for fields with drained periods [10].Much of the historical emissions of GHGs may be attributed to fossil fuel burning [11].Land-use change accounts for recent increases in emissions from fertilizer application, lowland rice fields due to fertilizer applied in water, and domestic animals such as cattle [12].
By far, agriculture and forest waste constitute the largest sources of GHG emissions from tropical countries such as Ghana [13].Agriculture and climate change are inextricably linked.Nelson [14] observed that "Agriculture is part of the climate change problem, contributing about 13.5% of annual GHG emissions (with forestry contributing an additional 19% compared with 13.1% from transportation)."Agriculture is, however, also part of the solution, offering promising opportunities for mitigation through carbon sequestration, improved soil and land-use management, and biomass production [14].e release of CO 2 from soil is the largest source of carbon emissions to the atmosphere [15].Soil CO 2 emissions and production are the result of complex interactions between climate and soil biological, chemical, and physical properties [16,17].
Soil surface CO 2 production is a major component of the biosphere's carbon cycle because it may constitute about three quarters of total ecosystem respiration [18].In recent years, soil CO 2 production has been the subject of intense studies because of its potential role in amplifying global warming [19].e rate of soil CO 2 production is dependent on land-use and land management systems [20].In Ghana, common land-use systems include forestry, upland agriculture, paddy rice, and animal husbandry [21].Understanding the controls on soil CO 2 emissions is critical because relatively small changes in soil CO 2 fluxes from these land-use systems may dramatically alter atmospheric concentrations of CO 2 .e critical factors reported to influence soil CO 2 production rates include atmospheric temperature and moisture, soil organic matter and nutrient content, root respiration, microbial processes, soil aeration, porosity and water, net primary productivity, and vegetation type [15,22].
For many years, most tropical countries such as Ghana have considered themselves as being net carbon sinks or, at worst, carbon neutral. is anecdotal assertion is based on the low level of industrialization in these countries.But given the extensive land-use change occurring in many tropical countries including deforestation and land degradation through poor management and periodic bush fires, it is conceivable that their GHG emissions are increasing [23].
ere are relatively few studies estimating GHG emissions in sub-Saharan West Africa, especially within the agricultural sector, and likewise, comparative studies across major landuse types are scarce.Consequently, the majority of practices and techniques for adaptation to climate change that are now being advocated [24,25] are largely based on knowledge generated in other parts of the world.e GHG inventory initiative of Ghana's Environmental Protection Agency (EPA) uses the Intergovernmental Panel on Climate Change (IPCC) guidelines to estimate GHG emissions from several sectors such as agriculture, forestry waste, animal manures, methane emissions from cattle, and lowland paddy rice fields [26].Findings from these estimates as well as those from the Carbon Dioxide Information Analysis Centre (CDIAC) (http: //www.cdiac.org)indicate that per capita carbon emissions in Ghana are on the increase.As stated by Milne et al. [27], a general weakness in these estimations is the heavy reliance on lower tier IPCC methodologies.Estimates by Ghana's EPA also show a gradual increase in GHG emissions with projected further increases based only on "best guesses" or by the use of emission factors (EFs) published by the IPCC [26].Actual measurements to validate these estimates or EFs are lacking.us, there is an urgent need for more assessments of ecosystem responses to land management (and mismanagement) in order to improve decision-making regarding climate change adaptation and mitigation. is study sought to address some of these identified knowledge gaps.It aims to measure the CO 2 emissions resulting from some of the major land-use systems operating within the coastal savanna agroecological zone of Ghana.

Study Area Description.
e CO 2 emissions experiment was carried out between July and November in 2012 at two locations with different land-use systems in the Coastal Savanna agroecological zone of Ghana.e first site, the Soil and Irrigation Research Centre (SIREC) at Kpong, University of Ghana, is located within the lower Volta basin (Figure 1).e 1,036 ha SIREC site is located at latitude 6 °09′ N and longitude 00 °04′ E, with an altitude of 22 m asl (Table 1).e second site, the Legon research farm, University of Ghana (main campus, Accra), is located at latitude 5 °66′ N and longitude 00 °19′ E, with an altitude of 88 m asl.e general 2 Applied and Environmental Soil Science topography of the SIREC-Kpong site is gently sloping with slopes ranging from 1 to 5%.e Legon-Accra site has a gentle, undulating relief with slopes ranging from 1 to 2%.
e Kpong site has an annual rainfall of 800-1326 mm, which is bimodal and characterised by a major rainy season (March-July), a short period of drought in August, a minor rainy season (September-November), and another period of drought (December-February) (Table 1).About 60% of the total rainfall occurs in the major rainy season and 30% in the minor rainy season.e rainfall distribution at the Legon site is similarly bimodal, with a mean annual rainfall range of 900-1010 mm.Prolonged heavy rain is occasionally experienced in the major rainy season from March/April to June whilst the minor rainy season begins from September/October to December.Temperatures at both study sites are warm.e mean maximum and minimum temperatures at the Kpong  site are 33.3 °C and 22.1 °C, respectively, and 30.9 and 21.5 °C, respectively, at the Legon site.e relative humidity for the night time to the early hours of the day for both sites ranges from 70 to 100%.e afternoon relative humidity ranges from 20% to 65% throughout the year.e vegetation at the Kpong site is limited to grasses and slow-growing, deep-rooting tree species. is is due to the soil's high clay content, and its shrink-swell characteristics and structure, combined with the climate effect (Table 1).
e main features of the natural vegetation in these soils are tolerance to drought, as well as development of deep roots to overcome root damage as a consequence of the annual cracking of the soil.e Legon site is covered with lush grass, thicket patches, and shrub vegetation community with little litter falls.Only a small amount of organic matter can therefore accumulate and the humus top soils are poorly developed.e soil at the Kpong site is an alluvial material derived from the weathering of garnetiferous hornblende gneiss (Table 1).It is classified as Typic calciustert [28].Locally, it is the tropical black clay called Akuse series [28] which is categorized as a vertisol [29].ese are generally deep black soils that contain more than 30% clay which is often dominated by smectite mineralogy [30].Generally, the clay content is very high in vertisols, and the dominant clay minerals are 2 : 1 type minerals (smectite and montmorillonites).At the Legon site, the soil is derived from a ferruginized weathered country rock, the Togo quartzite schists.It is classified as a ferric acrisol (sandy loam), which is a mineral soil with a characteristic argillic horizon [31].Locally, it is classified as a Toje series [32].

Experimental Layout for Sampling of Carbon Dioxide
Fluxes.Data were collected following a stratified random sampling approach.e sites sampled were stratified into land-use types and within each land-use type or strata, sampling for carbon dioxide was randomly done at three replicate locations.At the Legon site, the studied land-use systems were woodlot (Leucaena leucocephala), cultivated maize (Zea mays) field, and a natural forest stand.At the Kpong site, four land-use systems, namely, cultivated soya bean (Glycine max) field, natural forest, cattle kraal (an enclosure for cattle and other domestic animals), and lowland (paddy) rice field were considered (Figure 2).At the Kpong site, the soya bean was at the flowering stage.e soil did not receive any form of amendment (e.g., mineral fertilizer or manure).e land management system practiced includes ploughing with a tractor a week prior to sowing.e soya bean crops were under rainfed conditions throughout the growing season.e natural forest is at least 50 years old with the dominant tree species comprising Cassia fistula L., Ehretia anacua I. M. Johnst., and Azadirachta indica A. Juss.
e forest floor was covered by a thick mat of leaf litter and twigs.
e paddy field was under constant irrigation with about 5 cm head of flood water.At the time of sampling, the rice plants were at their emergence stage.
At the Legon site, the cultivated maize field was harvested prior to the sampling campaign.e field has been continuously under maize cultivation for more than five decades.Weeding is done by hand, and dead weeds and stovers from previous maize crops are left on the soil surface.e 20-year-old Leucaena leucocephala woodlot was adjacent to the cultivated maize field.Originally, this site was cultivated before its conversion to a woodlot for the production of fuelwood.e soil surface was covered with a thin layer of leaf litter.
e soil surface was covered by a thick mat of leaf litter and twigs.

Measuring Soil CO 2 Production.
e gas entrapment method described by Hutchinson and Mosier [33] and Sullivan et al. [34] was used.Transparent polyvinylchloride "PVC" chambers were inserted 2 cm into the mineral soil at the three random locations.A 10 ml solution of 3M NaOH was dispensed into a vial and placed under the plastic chamber to trap CO 2 evolving from the soil.Additional vials containing 10 ml of 3M NaOH placed in the transparent PVC with their lids on to exclude CO 2 evolved from the soil served as controls to account for the CO 2 trapped from the atmosphere.Measurement duration ranged from 9 to 15 days depending on the site.For each land-use system at the Kpong site, the trapping solutions were changed following these arrangements: (I) twice daily from the 19th to the 24th of July 2012 (12 h interval at 5:30 am and 5:30 pm for 6 days); (II) once daily from the 24th to the 27th of July 2012 (24 h interval at 5:30 pm for 4 days); and (III) once every two days from the 27th July to the 2nd of August (48 h interval at 5:30 pm for 7 days).For each landuse system at the Legon site, the trapping solutions were changed once daily (24 h) from the 28th of October to the 5th of November 2012 (9 days).e trapping chambers were placed at the same location after each measurement duration.After exposure of the alkali, the vials were removed, immediately covered with lids (air-tight seal), and taken to the laboratory for analysis.e evolved CO 2 was determined by back titration using a phenolphthalein indicator.

Soil Characteristics Sampling and Analyses.
e main soil characteristics with potential to influence CO 2 emissions were also measured.Prior to the beginning of the study, soil samples were taken by augering to a depth of 0-0.15 m at three random positions in each of the land-use systems at both study sites.Air-dried samples were bulked (for each land-use), crushed, and then sieved through a 2 mm sieve for characterization.e soil samples were analyzed for texture, pH, C, and N using the modified Bouyoucos hydrometer method [35], an electrode pH meter, the Walker and Black method, and the Kjeldahl method, respectively.
Soil temperature and soil moisture content were measured at the same time duration as gas sampling during the whole experimental period at the Legon site (only).Soil temperature was measured at a depth of 5 cm using a digital probe (pH/mV/C meter, RS232).Moisture content was determined by sampling with a core sampler and oven drying at 105 °C for 24 hours.Daily ambient air temperature and precipitation data (that can also influence soil temperature and moisture) were obtained from the weather data station at SIREC, Kpong.

Statistical Analysis.
e soil and environmental variables data were assessed using the dispersion and analysis of variance methods to relate differences to land-use systems [36].Analysis of variance was performed on soil CO 2 production rates on each sampling date separately, to assess differences between land-use systems and times during the day.Regression analysis was also used to determine the relationship between CO 2 production rates and environmental parameters (temperature and moisture) as expressed for each land-use system.To predict CO 2 production based on soil temperature, we used an exponential equation as suggested by Davidson et al. [37] and Raich & Potter [38].For soil water, we used a quadratic relationship between production and water content [37].Statistical differences were considered significant at p ≤ 0.05.In addition, the statistical package Statistix version 9.0 was used to test differences in means using the Tukey range test procedure at a significance level of p ≤ 0.05.Analysis of Applied and Environmental Soil Science variance was performed with Genstat statistical software (Genstat version 9.2).

Soil and Environmental
Variables.Table 2 summarizes information on the physical and chemical characteristics of the land-use systems at the Kpong and Legon sites, respectively, prior to commencement of the experiment.
e high clay content of the soils at the Kpong site con rms their vertic characteristic, whereas soils at the Legon site are predominantly sandy.e average pH of the Kpong site's vertisol soil is 7.0, described as neutral except for soils from the cattle kraal in which the pH was approximately 8.0 (alkaline).e Legon site's al sol soil is strongly acid.e organic carbon (OC) content di ered with each land-use system.e OC content of this site's kraal and forest soils is high.At the Legon site, the OC of the cultivated eld is low.e OC content of the forest oor is high (2.42%), whereas in the woodlot the OC is medium (1.55%).e total rainfall during the year of study (2012) at the Legon site was 594.7 mm (minor season), with only one small rainfall event (i.e., 5.1 mm) occurring during the measurement time frame.e average annual temperature in Kpong is 26.6 °C.e total rainfall at the Kpong site was 714 mm in the season where measurements were made.During the measurement time frame, ve rainfall events were recorded (i.e., 14.4, 46.6, 52.0, 3.6, and 0.5 mm), amounting to a total of 117.1 mm.Soil temperature varied between 28.95 and 36.6 °C during the study period at the Legon site (Figure 3(a)).
Temperatures were particularly high for the cultivated land-use system, whereas low soil temperatures were recorded in the forest land-use.Under the cultivated landuse, soil temperatures peaked during the second and fth sampling time and then decreased gradually to 34.7 °C.For the woodlot system, soil temperature increased gradually from 30.57°C to 32.7 °C during the rst and third sampling times.A sudden decrease in temperature then occurred on the sixth sampling time after which it up-surged to 35 °C and again decreased sharply to 29.37 °C.Low soil temperatures were found in the forest land-use, with a temperature average of 32.6 °C.A maximum temperature of 33.9 °C was measured during the fourth sampling time.e temperature then dipped to 28.95 °C during the last sampling time.
e moisture content of the Legon site's forest soils was relatively higher compared with the moisture contents of   Applied and Environmental Soil Science the cultivated and woodlot system soils (Figure 3(b)).e cultivated land-use recorded a low moisture content.e woodlot and cultivated eld initially recorded high moisture contents of 0.124 and 0.097 gg −1 , respectively, compared to 0.089 gg −1 from the forest soil.Moisture content decreased sharply to 0.065 and 0.038 gg −1 for the woodlot and cultivated eld, respectively, during the second sampling time.In most cases, woodlot soils stored much more moisture than cultivated soils.e moisture content of forest soils decreased gradually with time but was higher compared to the other land-use systems.

CO 2 Emission from a Clay Soil Environment (Kpong).
Soil CO 2 emissions di ered signi cantly with di erent landuse systems and for most measurement times.e highest CO 2 emission was observed from the cattle kraal, followed by the paddy rice and the forest ecosystem.Higher CO 2 uxes occurred during the daytime (5:30 am-5:30 pm) compared to emissions observed at night time (5:30 pm-5:30 am).
During the rst sampling time, the highest CO 2 emission of 340.5 mg•m −2 •h −1 was emitted from the kraal during the night time.During the day, the CO 2 production increased to 411.4 mg•m −2 •h −1 (Figure 4).e CO 2 emission pattern was maintained for sometime but decreased gradually to 226.Soil CO 2 emissions measured over a 24-hour interval were consistent with those based on a 12 h interval.For this period of measurement, cattle kraal CO 2 production was followed by emissions from the forest, whereas the paddy eld and cultivated land-uses emitted relatively lower CO 2 .Overall, during the whole measurement time, the highest average CO 2 emission was observed from the cattle kraal (256.7 mg•m −2 •h −1 ), followed by the forest (146.0 mg•m −2 •h −1 ) and paddy rice (140.6 mg•m −2 •h −1 ) land-uses.e lowest average emission was observed for the cultivated land (112.0 mg•m −2 •h −1 ).

CO 2 Emission from a Sandy Soil Environment (Legon).
Soil CO 2 emissions from the three land-use systems at the Legon site are shown in Figure 5.
Generally, low emissions were observed in the mornings, before peaking in the midafternoon and thereafter decreasing into the late afternoon (Figure 5(a)).In most cases, high CO 2 production was observed from the cultivated eld followed by emissions from the woodlot.Lower emissions were particularly recorded from the forest ecosystem (Figure 5(b)).
Soil CO 2 production from all of the land-use systems at rst sampling showed nonsigni cant di erences in emissions.
e average CO 2 production was 31.3 mg• m −2 •h −1 .CO 2 emissions ascended gradually at the second sampling time for  Applied and Environmental Soil Science both the forest and woodlot land-use systems.However, a steep increase in emissions was observed from the cultivated eld.A sudden drop in emissions to 37.6 mg•m −2 •h −1 was followed by a sharp increase to 88.0 mg•m −2 •h −1 which was the highest CO 2 production recorded for this land-use.Soil CO 2 emissions then dipped to 38.4 mg•m −2 •h −1 and then upsurged again to 78.0 mg•m −2 •h −1 where it nally declined to 45.9 mg•m −2 •h −1 at the last sampling time.
e CO 2 emissions from the woodlot showed a similar pattern as that of the cultivated eld, but the dynamics were gradual rather than steep.From 31.0 mg•m −2 •h −1 CO 2 during the rst sampling time, the CO 2 emissions increased gradually to 73.9 before decreasing sharply to 40.9 mg•m −2 •h −1 .ereafter, a gradual decrease and increase in emissions was maintained until a CO 2 production of 48.3 mg•m −2 •h −1 was recorded at the last sampling time.

Soil CO 2 Production, Temperature, and Moisture
Measurements on Sandy Soil Environment.A regression analysis reveals significant correlations between the respiration rate and soil temperature and moisture (p < 0.001).
e predictive power of the model, given by R 2 , was low in some cases.e regression of soil temperature on soil CO 2 production showed a positive correlation, with CO 2 evolution increasing as soil temperature increased (Figure 6).Soil temperature explained 65% of the total CO 2 production on cultivated land, 52% on woodland, and 29% on forest stand.Relationship between soil CO 2 production and volumetric soil moisture was higher in woodlot as compared with cultivated land and natural forest.

Discussion
4.1.Impacts of Land-Use Systems on CO 2 Emissions.Landuse and management practices may influence carbon inputs and hence CO 2 emissions [39].Indeed, the CO 2 emissions from different land-use systems at our study's Kpong site differed significantly.Higher CO 2 emissions were particularly observed from the cattle kraal and may be due to mineralization of this land-use's high organic matter content compared with the other land-use systems.Applications of organic manure to soil can increase CO 2 emissions [40].Indeed, after fresh organic matter input to soils, many specialized microorganisms grow quickly and to accelerate the soil organic matter leading to the priming effects [41].
McGill et al. [42] proposed that soluble organic C in the soil is an immediate source of C for soil microorganisms, which in turn emit CO 2 .Hence, large quantities of organic manure that are added to agricultural soils every year for supplying nutrients to crops may contribute significantly to CO 2 emission.
e measured organic matter content of the various land-use systems decreased in the order of kraal, forest, cropped land, and paddy rice.However, the initial high CO 2 emissions observed from the paddy rice field during the 12-hour sampling time could be due to adequate moisture content which increased microbial activity and hence enhanced the decomposition of organic matter.
ereafter, the emissions decreased steadily, and low CO 2 emissions were observed during the 24-and 48-hour measurement interval.
e onset of decreasing CO 2 production from the paddy rice field coincided with a period of flooding (irrigation) of the field.During this submerged period of paddy rice cultivation, CO 2 evolution in the soil is severely restricted due to the flooding condition [43].e soils of the studied forest land-uses contained a high amount of organic matter due to the accumulation of litter fall over time.During decomposition, microbial tissues and depolymerization products are produced which undergo chemical stabilization through complexation with mineral cations or physical stabilization by clays [44].Since vertisols contain heavy or high amounts of clay, the stabilized materials decompose about 100 times slower than the original litter [44].e forest soil CO 2 emission at the Kpong site was therefore low compared to emissions from the kraal.However, the emissions were significantly higher than emissions from the cultivated soils.
Cultivation of the soil increases the mineralization of the soil organic matter and hence the emission of CO 2 [45].
e decomposition of soil organic matter is increased by the physical disturbance caused by soil cultivation, which breaks down macroaggregates and exposes the carbon protected in their interiors to microbial processes [46].In this study, the low CO 2 emissions from the cultivated soil at the Kpong site could be partly due to its low organic matter content.Even though cultivation is expected to expose the organic matter to microbial decomposition, the heavy clay nature of this site's soil might have protected it.is may have significantly reduced the cultivated field's CO 2 emissions compared with the other land-use systems except for the paddy rice where flooding conditions impeded CO 2 emissions.At the Legon site, the cultivated field contained the lowest organic matter content, but it had high CO 2 emissions compared to the woodlot and forest land-uses.is may be due to the low clay content (i.e., sandy nature) of this site's alfisol soil which exposes the organic matter to microbial decomposition.
Soil temperature and moisture content are abiotic factors which influence processes that affect the dynamics of soil carbon.Soil microflora contributes 99% of the CO 2 arising as a result of decomposition of organic matter [47], while root respiration contributes 50% of the total soil respiration [48].Soil temperature affects microbial respiration, whereas soil moisture affects microbial respiration and soil respiration, and hence CO 2 evolution [49,50].Maximum CO 2 evolution was noted on the 1st and 3rd of November (at 88 and 78 mg•m −2 •h −1 , resp.). is may be attributed to the increasing role of root activity and organic matter decomposition in line with an increase in soil temperature which peaked at 36.5 and 35.7 °C on the 1st and 3rd of November, respectively.
At the Legon site, even though the forest floor had a higher organic matter content than that of the woodlot, low CO 2 emissions may be due to the low soil temperature slowing decomposition of its organic matter.Indeed, soil temperature can have a marked effect on CO 2 evolution from the soil [51].Considerable variations in soil CO 2 emissions during different periods (i.e., day and night) were observed.Soil CO 2 emissions from the various land-use systems during daytime were higher than the night time production.is may be attributed to the higher soil temperatures during the daytime measurements.

4.2.
Temperature and Moisture Effects on CO 2 Emissions.Soil water content and soil temperatures are known to be Applied and Environmental Soil Science important drivers of soil CO 2 production, and they may change as a result of forest thinning [52,53].Similar to Tang et al. [54], we used both soil water content and soil water content squared in our model.In many research studies, soil temperature was noted to be a strong and positive predictor of soil respiration, accounting for 43-75% of the variation in soil CO 2 production rates [55].On the other hand, increasing soil moisture would increase CO 2 evolution up to an optimum level, above which it would reduce CO 2 evolution [51].
e interaction of soil temperature and soil moisture assumes great significance in view of global warming and likely disturbance in precipitation patterns.However, Kowalenko et al. [56] observed that temperature was the most dominant factor in determining CO 2 evolution from the soil.
e regression of soil temperature on soil CO 2 production (Legon site) showed a positive correlation, with CO 2 evolution increasing as soil temperature increased.Soil temperature explained up to 65% (on cultivated land) of the total CO 2 production in the regression model.is strong relationship between soil temperature and CO 2 production is expected since soil respiration rates reflect heterotrophic and autotrophic activities that are highly temperature dependent [56]. is was reflected by the soil CO 2 emissions of the forest (with a low soil temperature) being low compared to the emissions from the kraal and cultivated land-use systems, of which the latter had a particularly high soil temperature.e temperature sensitivity coefficient (i.e., Q 10 values) is a convenient index for comparing the temperature sensitivity of soil CO 2 production.It is commonly used to express the relationship between soil biological activity and temperature [58].e Q 10 values from 25 to 35 °C for CO 2 emissions in this study suggests that CO 2 emission was controlled primarily by soil biological activity.It is estimated that a 1 °C increase in temperature could lead to a loss of 10% of soil organic carbon in regions of the world with an annual mean temperature of 25 °C [59].While in regions having a mean temperature of 30 °C, a 1 °C increase in temperature would lead to a 3% loss of soil organic carbon.

Conclusion and Way Forward
Measurement of CO 2 emissions from soils of different landuse systems allows the understanding and accurate evaluation of soil management practices to reduce GHG emissions.In our study, soil CO 2 emissions were significantly influenced by different land-use systems.Soil organic matter decomposition and mineralization were the main drivers of CO 2 emissions.e soil itself could serve as a source or sink of CO 2, depending on the management or land-use system imposed on it.Land-use systems which often disturb and expose the soil's organic matter to decomposition and mineralization are liable to emit more GHGs.
In our study, cattle kraals emitted large and increasing amounts of CO 2. is suggests that such kraals could become an increasing threat to global warming due to the large tracts of land occupied by livestock in developing countries.To reduce CO 2 emissions from cattle kraals, livestock management systems such as improved pasture with low stocking rates must be practiced.Our studied woodlot and forest land-uses recorded relatively low CO 2 emissions.is was despite the high organic matter content of their soils and could be attributed to the low level of soil disturbance in these land-uses.
is finding implies that maintaining forest reserves and promoting agroforestry systems that include woodlots is highly desirable for mitigating GHG emissions.We also found that CO 2 emissions from the lowland rice paddy field peaked when oxic conditions were maintained.Periodic flooding of the field (anoxic condition) often reduced CO 2 evolution; however, research studies show that this condition can promote CH 4 production.Due to the lack of access to a gas chromatograph (GC), other GHGs such as CH 4 and N 2 O could not be studied.While it is important to reduce CO 2 emissions through maintaining some head of water on the soil surface (i.e., flooding), periodic drainage is also important to reduce CH 4 emissions.
Overall, several factors influenced CO 2 emissions from the different land-use systems in our study.ese include inherent properties of the soils such as texture, temperature, and moisture content which influenced CO 2 production through their effect on soil microbial activity and root respiration.Soil temperature explained more than 50% of the variation in soil CO 2 production.A temperature coefficient sensitivity Q 10 of 4.1 depicts that the soil CO 2 emission was controlled primarily by soil microbial activity.
Hence, development and implementation of practices that increase tree cover to directly reduce emissions through carbon capture and sequestration should be of priority in the study area. is will help to mitigate global GHG emissions but importantly will also help to maintain or increase crop productivity and thereby improve global or regional food security.

Figure 1 :
Figure 1: Map of Ghana showing study areas: SIREC-Kpong and University of Ghana main campus, Legon-Accra.

Figure 2 :
Figure 2: Layout of identified land-use systems used for the study.

Figure 3 :
Figure 3: Variation of soil temperature (a) and soil moisture content (b) from di erent land-use systems at Legon farm.

6
3 mg•m −2 •h −1 during the fourth sampling time and up-surged to 421.3 mg•m −2 •h −1 during the fth sampling time during the day.Initially, the CO 2 emission from the paddy eld showed nonsigni cant di erences from the kraal.A CO 2 production of 330.0 mg•m −2 •h −1 was measured during the night time and increased to 404.3 mg•m −2 •h −1 during the day.e emission decreased gradually to 85.3 mg•m −2 •h −1 after which a sharp decrease resulted in a production of 31.3 mg•m −2 •h −1 .e forest and cultivated land-use systems initially revealed lower CO 2 emissions compared to the kraal and paddy eld but increased with time.e lowest CO 2 emission of 5.8 mg•m −2 •h −1 was from the forest land-use at the beginning of the measurement.is peaked to 112.8 mg•m −2 •h −1 during the daytime and dipped to 25.6 mg•m −2 •h −1 during the night time.Again, CO 2 emission ascended to 228.6 mg•m −2 •h −1 in the next sampling time and gradually decreased to 165.6 mg•m −2 •h −1 .e cultivated eld initially emitted 14.0 mg•m −2 •h −1 CO 2 , but this gradually increased to 176.6 mg•m −2 •h −1 , after which it decreased to 95.8 mg•m −2 •h −1 .e CO 2 production then up-surged to 198.8 mg•m −2 •h −1 and nally decreased to 84.2 mg•m −2 •h −1 .

Figure 6 :
Figure 6 : Relationship between soil water content and soil CO 2 production in di erent land-use systems on Ferric Acrisol at Legon farm, Coastal Savanna agroecological zone of Ghana.

Table 1 :
Main characteristics of the experimental sites.

Table 2 :
Initial soil chemical and physical properties of land-use systems at SIREC-Kpong (A) and Legon, University of Ghana (B).
Figure 4: Temporal CO 2 emission from di erent land-use at Kpong site.