Tropical Forest and REDD+ 2

The co-existence of trees, shrubs and grasses is a defining feature of the Tugai floodplain forest. During recent decades, the combined effect of drought and wetness has led to complex vegetation changes in the ecosystem. A number of recent earth observation based studies reported positive changes in biological productivity in the lower reaches of the Tarim region in relation to recurring discharges, triggering an increased amount of floodplain vegetation after the water conveyance periods. However, this ‘greening of the desert’ may only relate to deep-rooted species, with a potential reduction in grass cover having an important implication for the biodiversity composition.

The current study presents a large-scale earth observation based evaluation of long term trends in forest growth using the entire archive of Landsat images which is currently available for our study area. The Theil-Sen estimator served as trend indicator. The Google Earth Engine was used as processing platform which combines Google-scale storage and advanced cloud-based geospatial processing power in order to allow data driven remote sensing analysis.

The results suggest that the Poplar trees and Tamarisk shrubs are the main beneficiaries of the conveyance. The grasses can also benefit from the water, but only in very specific areas along the river. The results give restoration managers an important insight of the vegetation response and serve as decision support during the restoration project.

The Dry Forests biome is estimated to cover approximately 42% of sub-tropical and tropical forested regions in the world. In West-, East- and Southern Africa it covers approximately 17.3 million km²and is home to about 505 million people. Countries in the southern African regions such as Zambia, Zimbabwe and Malawi have some of the highest deforestation rates in Africa and globally; Malawi for example has an annual deforestation rate of 2.8% and is ranked 4th in the world and the first in southern Africa. For these reasons the southern African countries have all initiated national programmes in the UNFCCC REDD+ process. The countries require accurate, timely and regular forest monitoring systems to assess their deforestation and degradation rates via application of Earth Observation (EO). There are however several challenges for mapping these dry forest ecosystems. This includes the relatively low canopy closure on large areas with bare soils, shrubs and thickets visible through the tree canopy gaps which cause confusion of the spectral signals; this leads to higher uncertainty of the area assessments. Many of the species in the dry forests in Africa also are deciduous with leaf fall in the dry or winter season which poses a problem, as the optimal satellite images for the forest mapping can only be obtained in the winter season where the data is cloud free but trees are without leaves. During this season the forests have a very low canopy reflectance and therefore the mapping produces higher inaccuracies. The utilization of dense time series with multi-temporal imagery - which the Sentinel-2 data is well suited for – can solve problems caused by phenology changes of forest canopies between the seasons and data availability restrictions due to clouds. Furthermore, this data is also needed to monitor forest degradation which can be better detected by assessing forest canopy disturbances with high frequency time series. In the GSE FM REDD Extension project therefore a main objective was to develop methodologies based on the use of multi-temporal image data stacks and automated processing chains as a pre-cursor to the availability of the Sentinel-2 data for dry forest mapping. The methods were improved towards a full automatisation and a more robust processing chain. In order to simulate Sentinel-2 data, Landsat and SPOT5Take5 were used. The methods were successfully tested and validated in sites in Malawi and Zimbabwe with very high accuracy levels achieved of over 90% for forest mapping. These promising technical developments have been shared via capacity building with the countries. The scope for vastly improving the monitoring of the dry forest biome with Sentienl-2 data has implications for implementing international Conventions and/or programmes related to their conservation.

@font-face { font-family: "Calibri"; }p.MsoNormal, li.MsoNormal, div.MsoNormal { margin: 0cm 0cm 10pt; line-height: 115%; font-size: 11pt; font-family: Calibri; }.MsoChpDefault { font-size: 10pt; font-family: Calibri; }div.WordSection1 { page: WordSection1; }

In 2005, official discussions initiated at the United Framework Convention on Climate Change (UNFCCC) established a process to analyze how reducing greenhouse gas emissions from deforestation and degradation (REDD) in developing countries could contribute to the ultimate goal of the Convention. A basic underpinning requirement is the systematic long-term observation of forests and other land cover characteristics and changes on national and global levels. In response to this need several international activities have been established like the REDD working group within the “Global Observation of Forest and Land Cover Dynamics” (GOFC-GOLD) and the “Global Forest Observation Initiative” (GFOI) of GEO. Both initiatives are cooperating and foster a continuous interaction mechanism with the UNFCCC to provide technical contributions to the negotiations and for national-level capacity development for monitoring emphasizing the important role of satellite remote sensing in this context. Where GOFC-GOLD is addressing latest scientific findings in the REDD+ Source Book, GFOI provides with its Methods and Guidance Document hands on recipes for countries to build up operational Measurements, Reporting and Verification (MRV) systems.

The presentation will elaborate the state of play of REDD+ in the political context after COP-21 in Paris and discuss the approach and experiences of the Earth Observation and carbon monitoring community with the REDD political discussions. It addresses the research community with open scientific questions to improve methodologies to support REDD early actions and readiness mechanisms for building national REDD monitoring systems. Critical issues on monitoring forest degradation, accuracy assessments, sensor interoperability, forest biomass and emission factors (incl. biomass burning) and the role of evolving technologies will be discussed.

The tropical savannas cover around 20% of the earth’s land surface and while they have fewer trees and stored carbon than rainforests or temperate forests, their extent makes them significant in the global carbon cycle. In fact, Ciais et al. (2011) indicate a need to consider the contribution of savanna to the overall carbon budget. In Africa, savannas cover approximately 50% of the continent and are complex, heterogeneous, and open environments with high intra-species spectral variability due to differences in geology (e.g. granite and gabbro), topography, rainfall, herbivory and human impacts (e.g. fire, resource harvesting such as fuel wood or foliage browsing) within relatively short distances. The distribution of woody biomass in savannas reflects spatial patterns fundamental to such ecosystem processes.

In the frame of the ESA DUE GlobBiomass project, we aim at better estimating aboveground biomass (AGB) and its uncertainties by developing an innovative mapping approach in Eastern South Africa for the epochs 2005, 2010 and 2015, and forest change maps. This approach should contribute to the development of one global map for the year 2010. The South African 300 000 km² regional site is a North-South transect with a large precipitation gradient and various types of soil such as granit or gabbro. The regional site includes various forest types: savanna, enormous forest plantations stretching over hundreds of kilometers with various biomass classes of pine and eucalyptus forests, indigenous dense forests and coastal forests such as thickets. The regional site include Kruger National Park that is very well documented in term of disturbances such as fire and elephant effects.

CESBIO has developed a methodology to map AGB using L-band SAR data with an emphasis on the reduction of uncertainties, applied to woody ecosystems with AGB less than 150 t.ha^-1 (Mermoz et al., 2014). In this study, the method is refined and applied over South Africa using ALOS data for the year 2010, ALOS-2 and Sentinel-1 data for the year 2015. In situ AGB used in the method are measured by CSIR: 37 1-ha savanna plots measured in 2012 around the Kruger National Park and tens of 1-ha plots are being measuring in 2015 in savannas, plantations, dense forests and thickets.

The biomass maps developed in this study for the years 2010 and 2015 are used to assess the carbon stocks, and the loss of carbon when forests are disturbed. The biomass maps also permits to characterize the biomass change regimes in various landscapes such as the Kruger national park where savannas are mostly degraded by animals, and around the park where savannas suffer from anthropic pressure. Last but not least, the biomass maps allows to better understand how drivers such as climate, topography, soils, competition, herbivory, and fire impact the spatial savanna patterns, supporting studies such as Colgan et al. (2012).

REFERENCES:

Ciais, P., et al. "The carbon balance of Africa: synthesis of recent research studies." Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences 369.1943 (2011): 2038-2057

Mermoz, S., et al. "Biomass assessment in the Cameroon savanna using ALOS PALSAR data." Remote Sensing of Environment 155 (2014): 109-119.

Colgan, M. S., et al. "Topo-edaphic controls over woody plant biomass in South African savannas." Biogeosciences 9.5 (2012): 1809-1821.

The Cerrado is the second largest Brazilian biome that covers, with an extent of around 2 million km², approximately 25% of the national terrestrial surface. Characteristic for the Brazilian Cerrado is a gradient in vegetation structure and density, which is caused by variations in several environmental factors such as seasonality, soil fertility, long-term climatic fluctuations and fire events. This vegetation gradient forms heterogeneous landscape patterns from open grasslands to dense forests, which store large amounts of carbon in above- and below-ground biomass of the Cerrado vegetation. The ecosystem thus provides climate regulating services, which play an important role in the global carbon cycle. However, weak conservation laws and a rising global demand on agricultural areas put high pressure on the Cerrado. Currently, deforestation rates in the Cerrado even exceed those of Amazonia. It is therefore essential to reliably map the distribution and quantify the biomass in a reproducible manner to keep track of these processes and to support decision makers e.g. for climate change negotiations. Even though remote sensing data are a valuable source of information, there is still a lack of validated biomass maps that adequately address the structural complexity of the Cerrado. We therefore analysed the potential of Landsat time series and spaceborne multi-temporal hyperspectral (Hyperion) data for biomass quantification. We trained machine learning regression models based on spatially explicit biomass field data and remote sensing data to estimate the above-ground biomass. We analysed the potential of each remote sensing data set on its own as well as the combination of complementary multi- and hyperspectral data. This approach enabled us to estimate local above-ground biomass distribution and helped to identify the most important spectral and temporal features needed to quantify biomass changes over time. Based on these findings we can improve above-ground biomass quantification in the Cerrado and hopefully similar savanna-type ecosystems. This is specifically important considering forthcoming temporally and spectrally high resolution datasets, such as from Sentinel-2 and the Environmental Mapping and Analysis Program (EnMAP).