Remote Sensing in the Mekong Basin

Project Strategy Overview

      The objective of the current iresearch s to examine regional-scale landscape dynamics in river basins in Southeast Asia, relative to their connectivity to the South China Sea, with an emphasis on the Mekong River.  The geographic and geopolitical provenance of this project covers a diverse set of environments, being subject to rapid changes. Over the past year, it has become clear  that the Mekong River may be subject to rapid development of hydropower, where not only the upper "Chinese Cascade" of dams is in place, but that an additional 12 mainstem dams are under discussion, along with a projected 97 (!) on tributaries in Laos.  The cumulative effect of these dams would fundamentally alter the overall hydrologic and biogeochemical regime of the Mekong, (as well as fisheries and livelihoods) with consequences for the South China Sea. This project is in the position of both fundamental science challenges to understand what this might mean, as well as in a position of responsibility for not only analyzing the potential outcomes, but for making the results more broadly known.

    Our focus is on "high resolution but regional" work A basic premise is that the understanding of regional scale processes requires the higher resolution now possible with satellite data, process-based models, and field measurements, with the convergence of data and models from multiples sources. The work can be significantly extended, with the inclusion of information from specialized local and regional institutions. But such information is typically not readily accessible. A key regional player is the Mekong River Commission. Through a formal Letter of Agreement in place with the MRC, we are jointly developing a "Virtual Mekong Basin," to provide a portal to the world what the current and projected status is of the basin, under different conditions. This agreement opens the door further, to important data resources, as well as providing a portal for the use of NASA-information to broader communities.

      Our strategy has been to establish the overall climatology and hydrology of the broader region, then dialing down through to the biogeochemistry and ecosystem productivity of the Mekong and the Tonle Sap Lake, culminating in the hydrologic and chemical export to the sea. All of this has taken considerable time to assemble, with the not-unexpected hiccups along the way (especially for such a challenging geopolitical region). In the following sections we summarize thework on biogeochemistry.


      Rivers transport dissolved inorganic carbon (DIC) and organic carbon (OC) from land to the ocean, but they also chemically transform this carbon during its transit through in situ respiration and photosynthesis, and exchange CO2 with the atmosphere through gas exchange. Most rivers are supersaturated with CO2, and as such are a net source of carbon to the atmosphere on the order of 1 Pg C/year.  However, this value is poorly constrained and thought to be an underestimate. This CO2 supersaturation in rivers is driven by two processes, advection of CO2 produced by respiration in soil of their watersheds and/or hyporheic zones and the prevalence of net heterotrophic metabolism in rivers. Tropical rivers carry two-thirds of the global riverine organic carbon load, and many of them experience strong seasonal flooding, bringing large amounts of allochthonous carbon into the river. This high organic carbon load combined with high respiration rates in tropical climates creates the potential for high CO2 outgassing rates in tropical rivers.  

      Few studies have investigated seasonal and interannual variations in the pCO2 of large tropical rivers. We investigated the cumulative sequences of river metabolism organic carbon composition, and outgassing in the Mekong, as a large, tropical, carbonate-rich river influenced by a significant flood-drought cycle. Alin et al (2011) examined the physical controls on outgassing, with the all-important evaluation of the gas transfer velocity. An important parameter linking sediment transport to organic carbon is the changes in the composition of particulate organic matter (POM) over the course of the hydrograph. Ellis et al. (in review) examined the changes in the elemental and lignin compositions of POM, showing pronounced seasonal differences. Ellis et al. (in prep, near final) carried the compositional analysis further, to establish a baseline for terrestrial versus marine derived OM in coastal sediments, using a new proxy, the branched/isoprenoid tetraether (BIT) index. Finally, Ellis et al (draft) assessed the time scales over which terrestrial organic matter is exported from the Mekong watershed via fluvial processes, by looking at the ∆14C of the lignin phenols. Lockwood et al. (about to be submitted) established annual carbon budgets for the lower amazon, and the sequence of metabolic processes producing those budgets.

     Alin, S. R., M. F. F. L. Rasera, C. I. Salimon, J. E. Richey, G. W. Holtgrieve, A. V. Krusche, and A. Snidvongs (2011). Physical controls on carbon dioxide transfer velocity and flux in low‐gradient river systems and implications for regional carbon budgets, J. Geophys. Res., 116, G01009, doi:10.1029/2010JG001398

      Outgassing of carbon dioxide (CO2) from rivers and streams to the atmosphere is a major loss term in the coupled terrestrial‐aquatic carbon cycle of major low‐gradient river systems (the term "river system" encompasses the rivers and streams of all sizes that compose the drainage network in a river basin). However, the magnitude and controls on this important carbon flux are not well quantified. We measured carbon dioxide flux rates (FCO2), gas transfer velocity (k), and partial pressures (pCO2) in rivers and streams of the Amazon and Mekong river systems in South America and Southeast Asia, respectively. FCO2 and k values were significantly higher in small rivers and streams (channels <100 m wide) than in large rivers (channels >100 m wide). Small rivers and streams also had substantially higher variability in k values than large rivers. Observed FCO2 and k values suggest that previous estimates of basinwide CO2 evasion from tropical rivers and wetlands have been conservative and are likely to be revised upward substantially in the future. Data from the present study combined with data compiled from the literature collectively suggest that the physical control of gas exchange velocities and fluxes in low gradient river systems makes a transition from the dominance of wind control at the largest spatial scales (in estuaries and river mainstems) toward increasing importance of water current velocity and depth at progressively smaller channel dimensions upstream. These results highlight the importance of incorporating scale‐appropriate k values into basinwide models of whole ecosystem carbon balance.

     Ellis, E.E., R.G. Keil, A.I. Ingalls, and J.E.Richey. 2012. Seasonal variability in the sources of  particulate organic matter of the Mekong River as discerned by elemental and lignin analyses.   Journal of Geophysical Research-Biogeosciences. 117, GO1038, doi:101029/2011JG001816.

      The Mekong River ranks within the top ten rivers of the world in terms of water discharge and sediment load to the ocean, yet its organic matter (OM) composition remains unstudied. This river is experiencing anthropogenically-forced changes due to land use and impoundment, and these changes are expected to intensify in the future. Accordingly, we monitored bulk OM composition and vascular-plant signatures (using lignin phenols) of Mekong River particulate OM (POM) over a one-year period. Autochthonous production comprises a greater proportion of POM during the dry season than in the rainy season, as demonstrated by higher percent organic carbon values (7.9 ± 2.4 vs. 2.2 ± 0.4%), lower yields of lignin normalized to carbon (0.40 ± 0.05 vs. 1.1 ± 0.3 mg (100 mg OC)-1, and an increase in N:C ratios towards phytoplankton values during the dry season (from 0.06 to 0.11). Changes in the lignin-phenol composition of POM suggest that gymnosperms contribute more toward OM composition during the dry season, with angiosperms dominating in the wet season. This is supported by calculations of the lignin phenol vegetation index of riverine OM, which is statistically different among seasons (dry: 29.4 ± 15.6 vs. wet: 74.6 ± 15.6). These changes likely reflect seasonal differences in the proportion of flow that is coming from the Upper and Lower Basin, corresponding to compositional differences between the vegetation of these regions. Therefore, this work provides a baseline understanding of OM variability that can be used to assess how future change will affect this river.

     Ellis, E.E., A.I. Ingalls,  L.T. Truxal, R.G. Keil, and J.E.Richey (in prep. Draft in near-final revision). Sources and Temporal Variability of Branched and Isoprenoid Tetraether Lipids Exported by a Large Tropical River
      The organic carbon (OC) discharged by rivers is a key component of the global carbon cycle due to the potential for this material to be permanently buried in marine sediments. Accurate accounting of the proportion of terrestrial verses marine-derived organic matter preserved in coastal sediments is of considerable interest due to the paradox that less terrestrial carbon is preserved in ocean sediments than that which is predicted from riverine fluxes.  Recently, a new proxy has been developed to trace the proportion of terrestrially verses marine-derived organic matter in coastal sediments, known as the branched/isoprenoid tetraether (BIT) index. Although the BIT index has since been widely applied to assess the proportion autochthonous verses allochthonous carbon buried in marine sediments, the variability, and especially the seasonal variability, of the BIT index in the suspended sediment carried by rivers remains largely uninvestigated.  The objective of this study was to determine the variability of the BIT index of suspended sediments carried by the Mekong River, in Cambodia, just upstream of the delta, over the course of one year. Results showed that branched GDGTs can no longer be considered to be produced strictly in soils.  In this study, between 48 to 73% of the branched GDGTs associated with suspended and bed sediment of the Mekong River were intact, indicating that the cells from which these biomarkers are derived are viable.  Further, the percentage of intact branched GDGTs found in river bed and lake bed sediments, and suspended sediments generally exceed that found in soils.  Therefore, the significant in situ production of branched, in addition to isoprenoid, tetraether lipids suggests that aquatic environments substantially modify the original BIT index imparted on sediment in upland environments.

     Martin, E.E., A.E. Ingalls, J.E. Richey, R.G. Keil, L.T. Carlson, G.M. Santos, S.R. Alin, and E.R. M. Druffel. (2013). Age of riverine carbon suggests rapid export of terrestrial primary production in tropics. Geophysical Res. Letters 40, 1-5, doi:10.1002/2013GL057450, 2013
      The source and age of dissolved and particulate organic matter that is transported by the Mekong River provides dynamical information that bulk concentrations alone do not. We are assessing the time scales over which terrestrial organic matter is exported from the Mekong watershed via fluvial processes, and determining how seasonality affects the type of organic carbon being exported.  Lignin-derived phenols are a biomarker for vascular plants, as they are diagnostic of different plant material types and diagenetic condition, they are abundant, and they are fluvially transported. With a recently-developed technique, it is possible to radiocarbon date individual lignin phenols. Measurements were made from February 2009-February 2010, and analyzed for δ13C (organic matter source) and Δ14C (age), in the lower Mekong, near Phnom  Penh. Fine suspended sediments (FSS) and POC track each other and the hydrograph. DOC is enriched in ∆14C relative to POC, but depleted relative to lignin. Lignin results showed that plant-derived material that is characterizable as lignin is always modern..   The  ∆14C of POC is predictable and highly variable, with values ranging between -327.7 to +25.9‰, with the lowest values during low water and the highest at high water.  Such variability is consistent with the conclusion that organic carbon inputs are controlled by mountainous sources during high water and lowland and autochthonous sources during low water.

     Lockwood, D., J.E. Richey, P.D. Quay, M. Sampson, and M. Ung. (in prep, to be submitted November 2011). CO2 outgassing flux and ecosystem metabolism over an annual cycle in the Lower Mekong River  

      The main objectives of this paper are to (1) To estimate the CO2 degassing rate in the Lower Mekong River over an annual cycle. (2) To measure river metabolism over an annual cycle using the stable isotopic tracer δ18O, which has been used to estimate the metabolic state of river systems on integrated space and time scales. (3) To determine the processes controlling the CO2 degassing rate and observed seasonal cycles of biogeochemical variables using a combination of dissolved inorganic carbon (DIC) and O2 mass and isotope budgets. Measurements were made at a downstream site, by Phnom Penh, tracking the hydrographic seasons. Many of the biogeochemical indices we measured at the Phnom Penh site varied seasonally and in concert with the hydrograph.  DIC and alkalinity were both negatively correlated with discharge, however, the flux of DIC and alkalinity werepositively correlated with discharge, suggesting that weathering is enhanced by precipitation during the flood season.  pH was not significantly correlated with discharge, but notably, pH was highly correlated with the calcium carbonate (calcite) saturation index, suggesting that the geology of the basin strongly influences pH, particularly the carbonate source. However, for this site to be supersaturated throughout the entire year there must be a pervasive input of CO2 from soil or in situ respiration that keeps the water supersaturated. The seasonal trend in pCO2 opposes the alkalinity and DIC trends, peaking in the flood season and lowest in the dry season. O2 was undersaturated throughout our study, and O2 and CO2 were also inversely correlated. Together with the pCO2 supersaturation throughout the year, these observations suggest that respiration (or advection of soil CO2) causes the pCO2 supersaturation.  Using the dual isotopes of O2 and O2 saturation measurements, calculated R:P suggested that the Phnom Penh site was net heterotrophic during the entire year, and extremely heterotrophic during the flood season. Examining δ13C-DIC versus δ18O provides a view of which process controls river dissolved gas concentration at a certain time-photosynthesis, respiration or gas exchange. Calculations of the  δ13C of CO2 suggests that the CO2 outgassing from the Mekong River is very close to the average signature expected for terrestrial plant material, and is slightly heavier than the δ13C of FPOM.  The observation that δ13C of outgassing CO2 is -26‰ during the flood season suggests minimal influence of CO2 from the respiration of freshwater plankton; this supports our observation of a high R:P ratio and the possibility that P approaches zero at this time of year. CO2 outgassing fluxes are lower but on the same order of magnitude as flux measurements from the Amazon River basin. We calculated an annual flux of 5.4 Tg C. Although it is small compared to the outgassing contributed by the Amazon River (estimated at 0.5 Pg C/yr), this value is a significant on a global scale and is comparable to the DIC load of the Mekong River, which is 3.9 Tg C/yr. This estimate is most likely an underestimate of CO2 outgassing in the Lower Mekong Basin because we are not including outgassing from the Mekong delta or the Tonle Sap Great Lake in this calculation.

Remote Sensing and Modeling to Understand the Tonle Sap Great Lake

     The largest freshwater body in South East Asia, the Tonle Sap Lake (TSL) is a large fluvial lake and a unique ‘pulsing'ecosystem with a large floodplain, abundant biodiversity, and high seasonal sediment and nutrient fluxes from the Mekong River.  It is located in central Cambodia and features an unusual bi-directional hydraulic connection with the Mekong River. During the dry season, the TSL covers an area of about 2500 - 3500 km2 and drains into the Mekong via the Tonle Sap River (TSR). In the active southeast monsoon phase, the flow of the Tonle Sap River reverses as the Mekong swells with flood water, delivering more than 51,000 million m3 to the TSL. As a result, the lake expands more than five-fold flooding the surrounding alluvial plain and covering an area of about 14,500 - 16,000 km2.   The Mekong-Tonle Sap is also the most productive freshwater fishery in the world with 50 million people dependent on fisheries for nutrition, income, and cultural identity. Large-scale development in the upstream part of the Mekong may threaten the natural flow regime and reduce the flood pulse amplitude, impacting on the Tonle Sap ecosystem as well. Despite its importance, there is little basic ecological information about the TSL ecosystem and many scientific questions about the fishery remain unknown.

      Irvine et al (2011) established the overall chemical continuity between the Mekong River and the Tonle Sap. Kummu et al (in revision) developed a water balance model for the TSL, identifying the different water sources and outputs from the lake. Data from regional observations and from the3D-EIA model were used. An interesting aspect is that during the time of peak outflow from the TSL, when the Mekong flow is low, much of the water that transits downstream to the Delta and on to the ocean is actually from the TSL, with a totally different chemistry than the Mekong itself. (Hence any calculation of Mekong discharge to the South China Sea must include this aspect). The overall distributions of sediments in the TSL, and the consequences for ecosystem production, are very poorly known. Kirschke et al (in prep, near final) examined the spatial patterns of turbidity in different seasons, using spatial pattern metrics of MODIS 250 m data. They then related the remotely-sensed patterns to predictions from the 3D EIA model of suspended sediments and phytoplankton production (the same model used by Kummu et al). Finally, Holtgrieve et al. (in prep) evaluated long-term autonomous sensor data and dynamic Bayesian mass balance models to estimate rates of gross primary productivity (GPP) and ecosystem respiration (ER), with comparisons to the 3D-EIA model.

      K.N. Irvine, J.E. Richey, G.W. Holtgrieve, J. Sarkkula, and M. Sampson. 2011. Spatial and temporal variability of turbidity, dissolved oxygen, conductivity, temperature, and fluorescence in the lower mekong river-tonle sap system identified using continuous monitoring. International Journal of River Basin Management. DOI: 10.1080/15715124.2011.621430
     Continuous monitoring of turbidity, dissolved oxygen, conductivity, temperature, and fluorescence was done at five locations on the Tonle Sap Lake and the Mekong-Bassac Rivers near Phnom Penh, Cambodia, between 2004 and 2010 using autonomous datasondes. Seasonal, daily, and spatial trends were clearly identified in the data and were related to the annual monsoon rainy season-dry season cycle, system metabolism, system hydraulics, and in some cases, localized phenomena such as waste discharges. The datasondes were particularly useful to track the oxygenation of anoxic black water areas in the flooded forest fringe of the Tonle Sap that occurred during the rainy season freshwater pulse. A strongly developed vertical variation of turbidity, dissolved oxygen, and conductivity in the flooded forest fringe may be related to a combination of factors, including dissolved material release from bed sediment and a floating organics-rich particulate layer near the bottom of the lake. Grab samples for total suspended solids were collected at the Preak Leap site (Mekong River) in 2009 and 2010. An excellent relationship was established between daily mean turbidity and total suspended solids concentration for the Preak Leap site, with r2 = 0.95. ARIMA models adequately forecast water level and water quality data one month ahead.

      Kummu,M, S. Tes, S. Yin, P. Adamson, J. Józsa, J. Koponen, J. Richey,  and J. Sarkkula. 2013. Water balance analysis for the Tonle Sap Lake-floodplain system. Hydrol. Process. DOI: 10.1002/hyp.9718

     We present here a detailed water balance model, based on  observed data of discharges from the lake's tributaries, discharge between Mekong and the lake through the Tonle Sap River, precipitation, and evaporation. The overland flow between the Mekong and lake is modelled with EIA 3D hydrodynamic model. Over the eight-year simulation period, the model replicated the observed data rather well. There were six sub periods when the water balance model performed less well, particularly for the lower water levels occurring towards the end of the dry season during late April and early May. Breaking the overall water balance into its components shows the relative magnitude of each term. On average, 53.5% of Tonle Sap Lake volume originates from the Mekong mainstream, either via the Tonle Sap River (50%) or overland flow (3%). The system's own tributaries contribute 34%, and the balance of 12.5% is sourced from precipitation. The annual inflow during the eight-year study period ranged from 51 km3 during the dry conditions of 1998 to 109 km3 during the 2000 high flood episode. The estimated eight-year average inflow was 83.1 km3. Around 88.5% of the total annual outflow from the lake is discharged into the Mekong mainstream via the Tonle Sap River (84%) while the overland drainage back to the Mekong constitutes a fraction of 3%. The open water evaporative losses are assessed at 13%. The mean annual outflow during the study period was 81.9 km3, with annual volumes varying between 46 km3 (1998) and 114 km3 (2000). Flow alterations in the mainstream would have direct impacts on the Tonle Sap water levels and hydrology as well. Recent research has shown that the relatively small rises in the dry season lake water level would permanently inundate disproportionately large areas of floodplain, rendering it inaccessible to floodplain vegetation and eroding the productivity basis of the ecosystem by reducing the inundated area, and duration and amplitude of flooding. The lake extension would thus cause permanent submersion; in essence destruction, of considerable areas of the gallery forest stripe surrounding the lake in the floodplain. 

      Kirschke, S., J. E. Richey, M. P. Logsdon, and M. Kummu. (in prep).Variability of the spatial structure of turbidity in the Tonle Sap Lake , Cambodia. (Hydrology and Earth System Sciences).
     The flood pulse of the Mekong River carries dissolved and suspended solids to the lake and its floodplain, thus directly influencing the nutrient status of the flooded ecosystem.  But surprisingly few sediment data exist to assess system dynamics; precise sediment accumulation rates in the lake are unknown and estimates vary considerably. In Part 1 of this paper, we examine the seasonal variability in the spatial structure of turbidity. We use MODIS (Moderate Resolution Imaging Spectrometer) 250m level-1B satellite data for the TSL and an in situ data set of lake surface turbidity and total suspended sediments (TSS) to generate a turbidity index/product for the lake and analyze the variability in the spatial structure of turbidity between low, rising, high and falling water periods. The objectives are to (1) determine the correlation/relationship between the MODIS 250m band 1 and band 2 spectral band ratio and the in situ observational measurements of turbidity (NTU) and TSS; (2) calibrate the MODIS 250m derived turbidity index using the above mentioned relationship (spectral characteristics - NTU/TSS); (3) construct a data set of calibrated turbidity images for the TSL based on hydrographic stage data (low, rising, high, falling water stages); (4) identify spatial structure of turbidity in the data set by characterizing turbidity fronts on the turbidity (lake) surface; and (5) test the hypothesis that a relationship between the variability of turbidity and the spatial extent/stage height/time of year of the lake exists (on different scales) and can be analyzed using MODIS 250m satellite data. Geospatial metrics were used in the analysis, based on the Topographic Position Index (TPI). A series of metrics describing the TPI includes rugosity, number of patches, the perimeter-area fractal dimension, the largest patch index, and the splitting index illustrate seasonal differences in lake structure. In Part 2 of this paper, we compare the resulting spatial patterns to results from a model of TSS and phytoplankton distributions in the lake, as an independent and process-based evaluation of the observed patterns. Comparison to the EIA-3D model results, done with a reduced set of the spatial parameters, shows consistencies in patterns between the model and remotely sensed products, allowing a partitioning of net turbidity into suspended sediments and plankton.

     Holtgrieve, G.W, M.E. Arias, K.N. Irvine, D. Lamberts, E.J. Ward, M. Kummu, J. Koponen, J. Sarkkula,and J.E.Richey. (2013).  Patterns of ecosystem metabolism in the Tonle Sap Lake, Cambodia with links to capture fisheries. PLoS ONE 2013 Vol: 8(8):. DOI: 10.1371/journal.pone.0071395
      Numerous recent studies have highlighted the important role of freshwaters in the global carbon cycle (Cole et al. 1994, Richey et al. 2002). Mechanisms which increase the connectivity of inland waters to their surrounding watershed, for example, though annual flooding of riparian forests, are thought to highly influence internal ecosystem functioning of lakes and rivers through the supply of terrestrially derived organic matter (Cole et al. 2007). However, direct empirical evidence of geophysical processes controlling ecosystem processes is generally sparse. This project evaluated long-term autonomous sensor data and dynamic Bayesian mass balance models (Holtgrieve et al. 2010) to estimate rates of gross primary productivity (GPP) and ecosystem respiration (ER) in the Tonle Sap Lake over a full hydrologic cycle. There was a strong seasonal cycle in both the rates of GPP and ER that were correlated with lake water level.  High-water was associated with decreased GPP, elevated ER, and increasingly heterotrophic ecosystem dominated by respiratory process. This information is crucial for understanding the role of the Mekong-Tonle Sap Ecosystem in regional carbon sequestration. The results of our analyses are being integrated with previous modeling work to estimate the overall reliance of the fishery on in situ primary production.  How hydrology influences the energetic basis of these fisheries through fixation and respiration of organic matter is a particularly pressing question since current proposals for hydroelectric dam projects may threaten the long-term sustainability of the fishery by potentially removing the annual flood dynamics.