Land cover change and carbon emissions over 100 years in an African biodiversity hotspot.

Agricultural expansion has resulted in both land use and land cover change (LULCC) across the tropics. However, the spatial and temporal patterns of such change and their resulting impacts are poorly understood, particularly for the presatellite era. Here, we quantify the LULCC history across the 33.9 million ha watershed of Tanzania's Eastern Arc Mountains, using geo-referenced and digitized historical land cover maps (dated 1908, 1923, 1949 and 2000). Our time series from this biodiversity hotspot shows that forest and savanna area both declined, by 74% (2.8 million ha) and 10% (2.9 million ha), respectively, between 1908 and 2000. This vegetation was replaced by a fivefold increase in cropland, from 1.2 million ha to 6.7 million ha. This LULCC implies a committed release of 0.9 Pg C (95% CI: 0.4-1.5) across the watershed for the same period, equivalent to 0.3 Mg C ha(-1)  yr(-1) . This is at least threefold higher than previous estimates from global models for the same study area. We then used the LULCC data from before and after protected area creation, as well as from areas where no protection was established, to analyse the effectiveness of legal protection on land cover change despite the underlying spatial variation in protected areas. We found that, between 1949 and 2000, forest expanded within legally protected areas, resulting in carbon uptake of 4.8 (3.8-5.7) Mg C ha(-1) , compared to a committed loss of 11.9 (7.2-16.6) Mg C ha(-1) within areas lacking such protection. Furthermore, for nine protected areas where LULCC data are available prior to and following establishment, we show that protection reduces deforestation rates by 150% relative to unprotected portions of the watershed. Our results highlight that considerable LULCC occurred prior to the satellite era, thus other data sources are required to better understand long-term land cover trends in the tropics.

Correction to: Quantifying and understanding carbon storage and sequestration within the Eastern Arc Mountains of Tanzania, a tropical biodiversity hotspot.

Upon publication of the original article [1], the authors noticed that the figure labelling for Fig. 4 in the online version was processed wrong. The top left panel should be panel a, with the panels to its right being b and c. d and e should be the panels on the lower row, and f is correct. The graphs themselves are all correct. It is simply the letter labels that are wrong.

Quantifying and understanding carbon storage and sequestration within the Eastern Arc Mountains of Tanzania, a tropical biodiversity hotspot.

BACKGROUND: The carbon stored in vegetation varies across tropical landscapes due to a complex mix of climatic and edaphic variables, as well as direct human interventions such as deforestation and forest degradation. Mapping and monitoring this variation is essential if policy developments such as REDD+ (Reducing Emissions from Deforestation and Forest Degradation) are to be known to have succeeded or failed. RESULTS: We produce a map of carbon storage across the watershed of the Tanzanian Eastern Arc Mountains (33.9 million ha) using 1,611 forest inventory plots, and correlations with associated climate, soil and disturbance data. As expected, tropical forest stores more carbon per hectare (182 Mg C ha(-1)) than woody savanna (51 Mg C ha(-1)). However, woody savanna is the largest aggregate carbon store, with 0.49 Pg C over 9.6 million ha. We estimate the whole landscape stores 1.3 Pg C, significantly higher than most previous estimates for the region. The 95% Confidence Interval for this method (0.9 to 3.2 Pg C) is larger than simpler look-up table methods (1.5 to 1.6 Pg C), suggesting simpler methods may underestimate uncertainty. Using a small number of inventory plots with two censuses (n = 43) to assess changes in carbon storage, and applying the same mapping procedures, we found that carbon storage in the tree-dominated ecosystems has decreased, though not significantly, at a mean rate of 1.47 Mg C ha(-1) yr(-1) (c. 2% of the stocks of carbon per year). CONCLUSIONS: The most influential variables on carbon storage in the region are anthropogenic, particularly historical logging, as noted by the largest coefficient of explanatory variable on the response variable. Of the non-anthropogenic factors, a negative correlation with air temperature and a positive correlation with water availability dominate, having smaller p-values than historical logging but also smaller influence. High carbon storage is typically found far from the commercial capital, in locations with a low monthly temperature range, without a strong dry season, and in areas that have not suffered from historical logging. The results imply that policy interventions could retain carbon stored in vegetation and likely successfully slow or reverse carbon emissions.

Towards regional, error-bounded landscape carbon storage estimates for data-deficient areas of the world.

Monitoring landscape carbon storage is critical for supporting and validating climate change mitigation policies. These may be aimed at reducing deforestation and degradation, or increasing terrestrial carbon storage at local, regional and global levels. However, due to data-deficiencies, default global carbon storage values for given land cover types such as 'lowland tropical forest' are often used, termed 'Tier 1 type' analyses by the Intergovernmental Panel on Climate Change (IPCC). Such estimates may be erroneous when used at regional scales. Furthermore uncertainty assessments are rarely provided leading to estimates of land cover change carbon fluxes of unknown precision which may undermine efforts to properly evaluate land cover policies aimed at altering land cover dynamics. Here, we present a repeatable method to estimate carbon storage values and associated 95% confidence intervals (CI) for all five IPCC carbon pools (aboveground live carbon, litter, coarse woody debris, belowground live carbon and soil carbon) for data-deficient regions, using a combination of existing inventory data and systematic literature searches, weighted to ensure the final values are regionally specific. The method meets the IPCC 'Tier 2' reporting standard. We use this method to estimate carbon storage over an area of33.9 million hectares of eastern Tanzania, reporting values for 30 land cover types. We estimate that this area stored 6.33 (5.92-6.74) Pg C in the year 2000. Carbon storage estimates for the same study area extracted from five published Africa-wide or global studies show a mean carbon storage value of ∼50% of that reported using our regional values, with four of the five studies reporting lower carbon storage values. This suggests that carbon storage may have been underestimated for this region of Africa. Our study demonstrates the importance of obtaining regionally appropriate carbon storage estimates, and shows how such values can be produced for a relatively low investment.

Spatial clustering of porcine cysticercosis in Mbulu district, northern Tanzania.

BACKGROUND: Porcine cysticercosis is caused by a zoonotic tapeworm, Taenia solium, which causes serious disease syndromes in human. Effective control of the parasite requires knowledge on the burden and pattern of the infections in order to properly direct limited resources. The objective of this study was to establish the spatial distribution of porcine cysticercosis in Mbulu district, northern Tanzania, to guide control strategies. METHODOLOGY/PRINCIPAL FINDINGS: This study is a secondary analysis of data collected during the baseline and follow-up periods of a randomized community trial aiming at reducing the incidence rate of porcine cysticercosis through an educational program. At baseline, 784 randomly selected pig-keeping households located in 42 villages in 14 wards were included. Lingual examination of indigenous pigs aged 2-12 (median 8) months, one randomly selected from each household, were conducted. Data from the control group of the randomized trial that included 21 of the 42 villages were used for the incidence study. A total of 295 pig-keeping households were provided with sentinel pigs (one each) and reassessed for cysticercosis incidence once or twice for 2-9 (median 4) months using lingual examination and antigen ELISA. Prevalence of porcine cysticercosis was computed in Epi Info 3.5. The prevalence and incidence of porcine cysticercosis were mapped at household level using ArcView 3.2. K functions were computed in R software to assess general clustering of porcine cysticercosis. Spatial scan statistics were computed in SatScan to identify local clusters of the infection. The overall prevalence of porcine cysticercosis was 7.3% (95% CI: 5.6, 9.4; n = 784). The K functions revealed a significant overall clustering of porcine cysticercosis incidence for all distances between 600 m and 5 km from a randomly chosen case household based on Ag-ELISA. Lingual examination revealed clustering from 650 m to 6 km and between 7.5 and 10 km. The prevalence study did not reveal any significant clustering by this method. Spatial scan statistics found one significant cluster of porcine cysticercosis prevalence (P = 0.0036; n = 370). In addition, the analysis found one large cluster of porcine cysticercosis incidence based on Ag-ELISA (P = 0.0010; n = 236) and two relatively small clusters of incidence based on lingual examination (P = 0.0012 and P = 0.0026; n = 241). These clusters had similar spatial location and included six wards, four of which were identified as high risk areas of porcine cysticercosis. CONCLUSION/SIGNIFICANCE: This study has identified local clusters of porcine cysticercosis in Mbulu district, northern Tanzania, where limited resources for control of T. solium could be directed. Further studies are needed to establish causes of clustering to institute appropriate interventions.