I was really (and nerdily) excited after looking at the image below. A big, diffuse black spot over smooth forest, that isn’t there in the images taken 23 days before or after. Granted, I’m not 100% sure it’s a cloud, but what else could it be?

If you disagree, let me know in the comments!

Clouds in a Radarsat-1 image of Curuai Lake, Amazon floodplain, Brazil. Click for larger image.

If you’re curious about where in the world is this place, look here.

Usage of the image above is granted under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unreported License.

So far I have found three implementations of the test: mantel.rtest on the ade4 package, mantel on the ecodist package, and mantel on the vegan package.

Curious to see if there was any difference in processing speed,  I decided to time each implementation, with the same data:

> length(dst.bu04800) # first distance vector
[1] 405450

> length(dsg.bu04800) # second distance vector
[1] 405450

> # using ecodist
> system.time(man.bu.04800 <- mantel(dst.bu04800 ~ dsg.bu04800, nperm = 1000))
user  system elapsed
17.66    0.03   18.18

> # using ade4
> system.time(man.bu.04800 <- mantel.rtest(dst.bu04800,dsg.bu04800, nrepet = 1000))
user  system elapsed
39.33    5.99   45.71

> # using vegan
> system.time(man.bu.04800 <- mantel(dst.bu04800,dsg.bu04800, permutations = 1000))
user  system elapsed
53.07    4.42   57.69

For perspective, the machine I am using is a Dell Precision T7500 with Dual 2.13 Ghz Quad Core Xeon E5506 processors , and 6GB of RAM. I’m running R 2.13.0 64-bit, on Windows 7 Ultimate 64.

I guess that ecodist it is, at least for me.

But, just so you know, distance in ecodist is way slower than using dist in the base package:

> system.time(dst.bu04800 <- distance(bu04800$TOT_BU_AR))
user  system elapsed
13.49    0.02   13.51

> system.time(dst.bu04800 <- dist(bu04800$TOT_BU_AR))
user  system elapsed
0.01    0.00    0.02

If you think, for any reason, that I’m doing something wrong and there’s a faster way, let me know!

Abstract: Aquatic herbaceous macrophytes contribute significantly to the input carbon for the Amazon floodplain. These plants have large seasonal variations in areal coverage and high productivity. The present study estimates annual net primary production (NPP) of aquatic herbaceous macrophytes in a large lake on the eastern Amazon floodplain, assesses the sources and amount of uncertainty associated with these measures, and offers a comparison among the estimates of herbaceous macrophyte productivity in the Amazon region. Plant biomass accumulated during the rising water stage of the annual flood cycle, peaking at 2300 to 6100 g m-2 and decreasing later in the year. Annual net primary production was estimated to range from 2400 to 3500 g m-2 yr-1, with above water production between 650 and 1100 g m-2 yr-1, and below water production between 1700 and 2600 g m-2 yr-1. Echinochloa polystachya and Paspalum fasciculatum were the most productive species, followed by Paspalum repens, Hymenachne amplexicaulis, and Oryza perennis. The four main sources of uncertainty in the estimates were macrophyte taxa, location, sampling design, and lack of measurements of dead material loss.

To consult a statistician after an experiment is finished is often merely to ask him to conduct a post-mortem examination. He can perhaps say what the experiment died of.” (R. A. Fisher, 1938).

So true. That’s why someday I’ll find the time to get a second degree, in Stats.

Well, today I just found out about academia.edu, which is exactly that. The layout is great, simple and fast, and I can see the huge potential of it, once a good number of researchers is achieved. Give it a look, add yourself and find me there at http://uvic.academia.edu/ThiagoSilva. Finally, someone took the effort to bring social networking and the Web 2.0 to the world of academia!

Looking from above: Towards an integrative view of macrophyte growth and primary productivity in the Amazon floodplain.

Thiago Sanna F. Silva, Maycira M. P. F. Costa, John M. Melack

Aquatic macrophytes play important roles in the carbon budget of the Amazon floodplain due to their large coverage and high primary productivity, yet studies of macrophyte productivity in the Amazon region are few. The present study combines multi-temporal and multi-sensor satellite imagery and in situ data to develop a spatially and temporally consistent estimate of macrophyte coverage and primary production in the eastern Amazon floodplain. Field results indicate that plant biomass accumulates during rising water and reduces later in the season. Annual net primary production calculated from field data varied between 2470 and 3490 g/m2 (dry weight), with above water production between 650 and 1100 g/m2 and below water production between 1700 and 2570 g/m2. Remote sensing results indicate that within a single date, C-HV radar (ASAR) and infrared (MODIS) data have the highest potential for discriminating macrophytes from other cover types. The temporal variation in radiometric signature from macrophytes in C-HH at 35º (Radarsat) incidence angle and infrared (MODIS) differ the most from other cover classes. Further research is being conducted on the use of object-oriented segmentation/classification and spatial modeling to accurately map macrophyte cover and estimate macrophyte primary production over time, allowing for improvements in the current estimations of the contribution of macrophyte communities to the carbon budget of the Amazon floodplain.

Presentation (PDF, 3.5MB)

http://www.springerlink.com/content/w0528161605881w8/

The complete reference is:

Silva, Thiago Sanna Freire, Costa, M. P. F., Melack J. M. and Novo, E. M. L. M. Remote sensing of aquatic vegetation: theory and applications. Environmental Monitoring and Assessment 140:131-145. 2008.

doi: 10.1007/s10661-007-9855-3

(Poster – PDF)

Abstract:

RADAR REMOTE SENSING AS A TOOL TO ASSESS THE SEASONAL DYNAMICS OF MACROPHYTE COMMUNITIES IN THE AMAZON RIVER FLOODPLAIN

Silva, T. S. F., University of Victoria, Victoria, Canada.

Costa, M. P. F., University of Victoria, Victoria, Canada.

Melack, J. M., University of California at Santa Barbara, Santa Barbara, USA.

The Amazon floodplain participates significantly in the global carbon cycle. Macrophyte communities in this region display high productivity, rapid biomass turnover and the highest values of methane emission, representing a key element in the regional carbon balance. Their extent and inaccessibility, however, hinder the assessment of plant growth dynamics. Radar remote sensing presents a valuable tool in these conditions, allowing monthly spatial data acquisition. Radar backscatter is associated with plant structural characteristics, and radar imagery has been used with success to estimate cover and biomass of plant stands. For the present study, monthly Radarsat and Envisat ASAR imagery were acquired from December/03 to November/04, together with plant biomass field data. Variation in macrophyte radar backscatter corresponded to the seasonal pattern of water level variation, which in turn determines macrophyte growth patterns. Biomass and backscatter relationship is being further evaluated, in order to utilize image classification and modelling techniques to estimate macrophyte cover, stand biomass and productivity. This knowledge is essential in order to improve the accuracy of current ecosystem models and carbon emission and cycling estimates in the Amazon floodplain.

TS-C07 TS-B03 TS-C04

(pdf)

VARIAÇÃO SAZONAL DA RESPOSTA RADIOMÉTRICA DE COMUNIDADES DE MACRÓFITAS AMAZÔNICAS EM IMAGENS RADARSAT E ENVISAT ASAR.

Thiago Sanna F. Silva (thiago@uvic.ca), Maycira P. F. Costa

Department of Geography – University of Victoria
PO BOX 3050 STN CSC – Victoria, B.C., V8W 3P5, Canada.

As áreas alagáveis (wetlands) compreendem aproximadamente 4 a 6% da superfície terrestre, ocorrendo em todos climas e continentes (Mitsch et al., 1994). Estas áreas apresentam participação considerável no ciclo global de carbono, sendo responsáveis por até 72% das emissões de metano por fontes não-antropogênicas (Wuebbles e Hayhoe, 2002). A planície de inundação do rio Amazonas é uma das mais extensas áreas alagáveis do planeta, e desta forma, responsável por uma contribuição significante ao fluxo global de metano para atmosfera (Melack et al., 2004). Atualmente, estimativas mais precisas destes fluxos na região são limitadas por incertezas na dinâmica sazonal e interanual das áreas inundadas. Sabe-se, contudo, que áreas de crescimento de plantas aquáticas (macrófitas) apresentam alguns dos mais altos valores de emissão de metano (Devol et al,.1990). Neste contexto, dados obtidos por sensores remotos apresentam-se como uma das ferramentas mais adequadas no estudo da dinâmica das comunidades de macrófitas. Dentre estes, podemos destacar os sensores por radar, capazes de imagear a planície amazônica independemente de cobertura nebulosa ou iluminaçao solar. No presente estudo, serão utilizadas imagens RADARSAT e ENVISAT ASAR para a avaliação da resposta radiométrica de macrófitas, em duas localidades próximas à cidade de Santarém-PA: Lago Grande de Monte Alegre e Lago Grande de Curuai. Tais imagens foram adquiridas em diferentes datas durante os anos de 2003 e 2004, nas seguintes configurações: ângulo de incidência de 27º, banda C, polarização HH (RADARSAT, ENVISAT) e VV (ENVISAT). Concomitantemente com a aquisição de imagens, foram realizadas amostragens em campo descrevendo temporalmante a composição específica e biomassa de plantas aquáticas emersas. As imagens adquiridas foram submetidas à calibraçao radiométrica e geométrica, a fim de permitir comparações entre diferentes datas, sensores e dados terrestres. A partir das imagens calibradas, utilizando-se dos dados de campo e interpretação visual, serão identificadas áreas de ocorrência de plantas aquáticas. Nestes locais, serão extraídos os valores de retorno de radiação (Backscatter) nas diferentes datas e imagens, os quais serão utilizados para quantificação e caracterização da variação na resposta radiométrica exibida por tais comunidades. As séries temporais serão então empregadas na otimização de algoritmos de classificação de imagens. Será também avaliada a relação entre radiometria e a variação sazonal de biomassa destas comunidades, objetivando o desenvolvimento de modelos de produtividade primária. Estudos prévios indicam uma variação sazonal de até 5db na resposta radiométrica associada à vegetação aquática, a partir de imagens Radarsat, com máximo durante a estação de cheias (Costa, 2000). Contudo, o mesmo estudo evidencia a variabilidade deste comportamento de acordo com as diferentes configurações de imageamento. Estudos realizados anteriormente acerca deste tópcio limitam-se a avaliar perídos isolados, usualmente coincidindo com situaçoes de maior e menor nível de inundação. O presente estudo representa a primeira tentaviva em se caracterizar a resposta radiométrica de comunidades de plantas aquáticas em um ciclo sazonal completo, com frequência temporal de aproximadamente 30 dias.

Referências

Costa, M. 2000. Net Primary Productivity of Aquatic Vegetation of the Amazon Floodplain: a multi-SAR Satellite Approach. Phd Thesis, Department of Geography, University of Victoria, Canada.

Devol, A. H., Richey, J. Forsberg, B., Martinelli, L.A. Seasonal dynamics in methane emissions from the Amazon River floodplain to the troposphere. Journal of Geophysical Research 95(D10): 16417-16426. 1990

Melack, J. M., Hess, L. L., Gastil, M., Forsberg, B. R., Hamilton, S. K., Lima, I. B. T., Novo, E. M. L. M. Regonalization of Methane Emissions in the Amazon Basin with Microwave Remote Sensing. Global Change Biology. 2004.

Mitsch, W. J., Mitsch, R. H., Turner, R. E. Wetlands of the Old and New Worlds: ecology and management. In: Global Wetlands: Old World and New. Elsevier, 1994.